Merge branch 'main' into staging

.gitea/workflows/build.yml

@@ -15,6 +15,7 @@ env:
   WEB_IMAGE: apps/myai-web
   CV_CLEANUP_JOB_IMAGE: apps/myai-cv-cleanup-job
   CV_SEARCH_JOB_IMAGE: apps/myai-cv-search-job
+  PAGE_FETCHER_API_IMAGE: apps/myai-page-fetcher-api
   IMAGE_TAG: staging
 
 jobs:
@@ -62,6 +63,10 @@ jobs:
         run: |
           docker build -f Jobs/cv-search-job/Dockerfile -t "${REGISTRY_HOST}/${CV_SEARCH_JOB_IMAGE}:${IMAGE_TAG}" .
 
+      - name: Build Page Fetcher API image
+        run: |
+          docker build -f Apis/page-fetcher-api/Dockerfile -t "${REGISTRY_HOST}/${PAGE_FETCHER_API_IMAGE}:${IMAGE_TAG}" .
+
       - name: Push API image
         run: |
           docker push "${REGISTRY_HOST}/${API_IMAGE}:${IMAGE_TAG}"
@@ -88,4 +93,8 @@ jobs:
 
       - name: Push CV search job image
         run: |
-          docker push "${REGISTRY_HOST}/${CV_SEARCH_JOB_IMAGE}:${IMAGE_TAG}"
+          docker push "${REGISTRY_HOST}/${CV_SEARCH_JOB_IMAGE}:${IMAGE_TAG}"
+
+      - name: Push Page Fetcher API image
+        run: |
+          docker push "${REGISTRY_HOST}/${PAGE_FETCHER_API_IMAGE}:${IMAGE_TAG}"

Apis/api/Program.cs

@@ -5,8 +5,8 @@ using Email.Data;
 using Email.Data.Repositories;
 using Email.Data.Repositories.Contracts;
 using Email.Data.Services;
-using EmailApi.Models.Clients;
-using EmailApi.Models.Settings;
+using Email.Models.Clients;
+using Email.Models.Settings;
 using Microsoft.EntityFrameworkCore;
 using Models.Settings;
 using MyAi.Data;

Apis/api/Services/EmailApiEmailSender.cs

@@ -1,8 +1,8 @@
 using Api.Services.Contracts;
 using CvMatcher.Models.Responses;
 using Email.Data.Services;
-using EmailApi.Models.Clients;
-using EmailApi.Models.Requests;
+using Email.Models.Clients;
+using Email.Models.Requests;
 using Microsoft.Extensions.Options;
 using Models.Requests;
 using Models.Settings;

Apis/common/Settings/PageFetcherApiSettings.cs

@@ -0,0 +1,11 @@
+namespace Common.Settings;
+
+/// <summary>
+/// Connection settings for the internal page-fetcher-api service.
+/// Bound from the <c>PageFetcherApi</c> configuration section.
+/// </summary>
+public sealed class PageFetcherApiSettings
+{
+    public string BaseUrl { get; set; } = string.Empty;
+    public string InternalApiKey { get; set; } = string.Empty;
+}

Apis/cv-matcher-api/Dockerfile

@@ -8,6 +8,7 @@ COPY Apis/cv-search-data/cv-search-data.csproj Apis/cv-search-data/
 COPY Apis/cv-matcher-data/cv-matcher-data.csproj Apis/cv-matcher-data/
 COPY Apis/common/common.csproj Apis/common/
 COPY Apis/cv-matcher-api-models/cv-matcher-api-models.csproj Apis/cv-matcher-api-models/
+COPY Apis/page-fetcher-api-models/page-fetcher-api-models.csproj Apis/page-fetcher-api-models/
 COPY Apis/myai-data/myai-data.csproj Apis/myai-data/
 COPY Apis/shared-data/shared-data.csproj Apis/shared-data/
 COPY Helpers/common-helpers/common-helpers.csproj Helpers/common-helpers/
@@ -20,6 +21,7 @@ COPY Apis/cv-search-data/ Apis/cv-search-data/
 COPY Apis/cv-matcher-data/ Apis/cv-matcher-data/
 COPY Apis/common/ Apis/common/
 COPY Apis/cv-matcher-api-models/ Apis/cv-matcher-api-models/
+COPY Apis/page-fetcher-api-models/ Apis/page-fetcher-api-models/
 COPY Apis/myai-data/ Apis/myai-data/
 COPY Apis/shared-data/ Apis/shared-data/
 COPY Helpers/common-helpers/ Helpers/common-helpers/

Apis/cv-matcher-api/Program.cs

@@ -13,6 +13,7 @@ using Microsoft.EntityFrameworkCore;
 using Refit;
 using Serilog;
 using Common.Settings;
+using PageFetcher.Models;
 using StartupHelpers;
 using System.Reflection;
 
@@ -36,6 +37,16 @@ try
     builder.Services.Configure<CvMatcher.Models.Settings.AiSettings>(builder.Configuration.GetSection("Ai"));
     builder.Services.Configure<MatcherSettings>(builder.Configuration.GetSection("Matcher"));
     builder.Services.Configure<JobSearchSettings>(builder.Configuration.GetSection("JobSearch"));
+    builder.Services.Configure<PageFetcherApiSettings>(builder.Configuration.GetSection("PageFetcherApi"));
+
+    builder.Services.AddRefitClient<IPageFetcherApiClient>()
+        .ConfigureHttpClient((sp, c) =>
+        {
+            var settings = sp.GetRequiredService<Microsoft.Extensions.Options.IOptions<PageFetcherApiSettings>>().Value;
+            c.BaseAddress = new Uri(settings.BaseUrl.TrimEnd('/') + "/");
+            if (!string.IsNullOrWhiteSpace(settings.InternalApiKey))
+                c.DefaultRequestHeaders.Add("X-Internal-Api-Key", settings.InternalApiKey);
+        });
 
     builder.Services.AddRefitClient<IRefitRagApi>()
         .ConfigureHttpClient((sp, c) =>
@@ -50,7 +61,7 @@ try
 
     builder.Services.AddScoped<IRagApiClient, RagApiClient>();
     builder.Services.AddHttpClient<IMatcherAiClient, MatcherAiClient>();
-    builder.Services.AddHttpClient<IJobTextExtractor, JobTextExtractor>();
+    builder.Services.AddScoped<IJobTextExtractor, JobTextExtractor>();
 
     builder.Services.AddDbContext<CvMatcherDbContext>(options =>
     {

Apis/cv-matcher-api/Services/JobTextExtractor.cs

@@ -1,26 +1,23 @@
-using System.Net;
-using System.Text.RegularExpressions;
 using CvMatcher.Models.Settings;
 using Api.Services.Contracts;
 using Microsoft.Extensions.Options;
+using PageFetcher.Models;
 
 namespace Api.Services;
 
 /// <summary>
 /// Extracts normalised plain text from a job posting, either from a pasted description or by
-/// fetching and stripping the HTML of the job page URL.
+/// fetching the job page text via <c>page-fetcher-api</c> (headless Chromium rendering).
 /// </summary>
 public sealed class JobTextExtractor : IJobTextExtractor
 {
-    private readonly HttpClient _http;
+    private readonly IPageFetcherApiClient _pageFetcher;
     private readonly MatcherSettings _settings;
 
-    public JobTextExtractor(HttpClient http, IOptions<MatcherSettings> options)
+    public JobTextExtractor(IPageFetcherApiClient pageFetcher, IOptions<MatcherSettings> options)
     {
-        _http = http;
+        _pageFetcher = pageFetcher;
         _settings = options.Value;
-        _http.Timeout = TimeSpan.FromSeconds(25);
-        _http.DefaultRequestHeaders.UserAgent.ParseAdd("MyAi.ro CV Matcher/1.0");
     }
 
     /// <inheritdoc />
@@ -31,15 +28,18 @@ public sealed class JobTextExtractor : IJobTextExtractor
 
         if (string.IsNullOrWhiteSpace(jobUrl)) return string.Empty;
         if (!Uri.TryCreate(jobUrl, UriKind.Absolute, out var uri) || uri.Scheme is not ("http" or "https"))
-        {
             throw new InvalidOperationException("Invalid job URL.");
-        }
 
-        var html = await _http.GetStringAsync(uri, ct);
-        html = Regex.Replace(html, "<script[\\s\\S]*?</script>", " ", RegexOptions.IgnoreCase);
-        html = Regex.Replace(html, "<style[\\s\\S]*?</style>", " ", RegexOptions.IgnoreCase);
-        html = Regex.Replace(html, "<[^>]+>", " ");
-        return Limit(Normalize(WebUtility.HtmlDecode(html)));
+        var response = await _pageFetcher.FetchAsync(new FetchPageRequest
+        {
+            Url = jobUrl,
+            CallerService = "cv-matcher-api"
+        }, ct);
+
+        if (!response.Success)
+            throw new InvalidOperationException($"Failed to fetch job page: {response.Error}");
+
+        return Limit(Normalize(response.Text));
     }
 
     /// <summary>Truncates text to the configured maximum character count.</summary>

Apis/cv-matcher-api/cv-matcher-api.csproj

@@ -82,6 +82,7 @@
     <ProjectReference Include="..\cv-search-data\cv-search-data.csproj" />
     <ProjectReference Include="..\cv-matcher-data\cv-matcher-data.csproj" />
     <ProjectReference Include="..\common\common.csproj" />
+    <ProjectReference Include="..\page-fetcher-api-models\page-fetcher-api-models.csproj" />
     <ProjectReference Include="..\..\Helpers\startup-helpers\startup-helpers.csproj" />
   </ItemGroup>
 </Project>

Apis/email-api-models/Clients/IEmailApiClient.cs

@@ -1,7 +1,7 @@
-using EmailApi.Models.Requests;
+using Email.Models.Requests;
 using Refit;
 
-namespace EmailApi.Models.Clients;
+namespace Email.Models.Clients;
 
 public interface IEmailApiClient
 {

Apis/email-api-models/Requests/SendEmailRequest.cs

@@ -1,4 +1,4 @@
-namespace EmailApi.Models.Requests;
+namespace Email.Models.Requests;
 
 public sealed class SendEmailRequest
 {

Apis/email-api-models/Settings/EmailApiSettings.cs

@@ -1,4 +1,4 @@
-namespace EmailApi.Models.Settings;
+namespace Email.Models.Settings;
 
 public sealed class EmailApiSettings
 {

Apis/email-api-models/Settings/SmtpSettings.cs

@@ -1,4 +1,4 @@
-namespace Models.Settings;
+namespace Email.Models.Settings;
 
 public sealed class SmtpSettings
 {

Apis/email-api/Controllers/EmailController.cs

@@ -1,9 +1,9 @@
-using EmailApi.Models.Requests;
-using EmailApi.Services;
+using Api.Services;
+using Email.Models.Requests;
 using Microsoft.AspNetCore.Mvc;
 using Swashbuckle.AspNetCore.Annotations;
 
-namespace EmailApi.Controllers;
+namespace Api.Controllers;
 
 /// <summary>
 /// Internal email relay. Accepts an HTML body fragment from trusted callers

Apis/email-api/Program.cs

@@ -3,8 +3,9 @@ using Email.Data;
 using Email.Data.Repositories;
 using Email.Data.Repositories.Contracts;
 using Email.Data.Services;
-using EmailApi.Services;
+using Api.Services;
 using Microsoft.EntityFrameworkCore;
+using Email.Models.Settings;
 using Models.Settings;
 using Serilog;
 using StartupHelpers;

Apis/email-api/Services/SmtpEmailDispatcher.cs

@@ -1,12 +1,13 @@
 using Email.Data.Services;
-using EmailApi.Models.Requests;
+using Email.Models.Requests;
 using MailKit.Net.Smtp;
 using MailKit.Security;
 using Microsoft.Extensions.Options;
 using MimeKit;
+using Email.Models.Settings;
 using Models.Settings;
 
-namespace EmailApi.Services;
+namespace Api.Services;
 
 /// <summary>
 /// Wraps an HTML body fragment in the branded HTML shell and sends the resulting email via SMTP using MailKit.

Apis/page-fetcher-api-models/FetchPageRequest.cs

@@ -0,0 +1,20 @@
+namespace PageFetcher.Models;
+
+/// <summary>
+/// Request to fetch a web page via the page-fetcher-api.
+/// </summary>
+public sealed class FetchPageRequest
+{
+    /// <summary>Absolute HTTP or HTTPS URL to fetch.</summary>
+    public string Url { get; set; } = string.Empty;
+
+    /// <summary>
+    /// Playwright wait condition. Accepted values: <c>networkidle</c> (default), <c>domcontentloaded</c>, <c>load</c>.
+    /// </summary>
+    public string WaitFor { get; set; } = "networkidle";
+
+    /// <summary>
+    /// Identifies the calling service for audit purposes (e.g. <c>cv-matcher-api</c>, <c>cv-search-job</c>).
+    /// </summary>
+    public string CallerService { get; set; } = string.Empty;
+}

Apis/page-fetcher-api-models/FetchPageResponse.cs

@@ -0,0 +1,25 @@
+namespace PageFetcher.Models;
+
+/// <summary>
+/// Result of a page fetch operation.
+/// </summary>
+public sealed class FetchPageResponse
+{
+    /// <summary>Final URL after any redirects.</summary>
+    public string Url { get; set; } = string.Empty;
+
+    /// <summary>HTTP status code returned by the page. <c>0</c> on network failure.</summary>
+    public int StatusCode { get; set; }
+
+    /// <summary>Full rendered HTML as returned by Playwright.</summary>
+    public string Html { get; set; } = string.Empty;
+
+    /// <summary>Plain text extracted from the HTML (script/style stripped, whitespace normalised).</summary>
+    public string Text { get; set; } = string.Empty;
+
+    /// <summary>Whether the fetch succeeded. <c>false</c> on timeout or network error.</summary>
+    public bool Success { get; set; }
+
+    /// <summary>Exception message when <see cref="Success"/> is <c>false</c>.</summary>
+    public string? Error { get; set; }
+}

Apis/page-fetcher-api-models/IPageFetcherApiClient.cs

@@ -0,0 +1,16 @@
+using Refit;
+
+namespace PageFetcher.Models;
+
+/// <summary>
+/// Refit client for the internal page-fetcher-api service.
+/// All calls require the <c>X-Internal-Api-Key</c> header, configured at registration time.
+/// </summary>
+public interface IPageFetcherApiClient
+{
+    /// <summary>
+    /// Fetches a web page via headless Chromium and returns the rendered HTML and extracted plain text.
+    /// </summary>
+    [Post("/api/page/fetch")]
+    Task<FetchPageResponse> FetchAsync([Body] FetchPageRequest request, CancellationToken ct = default);
+}

Apis/page-fetcher-api-models/Settings/PageFetcherSettings.cs

@@ -0,0 +1,17 @@
+namespace PageFetcher.Models.Settings;
+
+/// <summary>
+/// Runtime settings for the page-fetcher service.
+/// Bound from the <c>PageFetcher</c> configuration section.
+/// </summary>
+public sealed class PageFetcherSettings
+{
+    /// <summary>Default Playwright wait condition (<c>networkidle</c>, <c>load</c>, <c>domcontentloaded</c>).</summary>
+    public string DefaultWaitFor { get; set; } = "networkidle";
+
+    /// <summary>Page navigation timeout in seconds.</summary>
+    public int TimeoutSeconds { get; set; } = 30;
+
+    /// <summary>Maximum characters stored/returned in the extracted text field.</summary>
+    public int MaxTextChars { get; set; } = 60_000;
+}

Apis/page-fetcher-api-models/page-fetcher-api-models.csproj

@@ -0,0 +1,12 @@
+<Project Sdk="Microsoft.NET.Sdk">
+  <PropertyGroup>
+    <TargetFramework>net10.0</TargetFramework>
+    <Nullable>enable</Nullable>
+    <ImplicitUsings>enable</ImplicitUsings>
+    <AssemblyName>page-fetcher-api-models</AssemblyName>
+    <RootNamespace>PageFetcher.Models</RootNamespace>
+  </PropertyGroup>
+  <ItemGroup>
+    <PackageReference Include="Refit.HttpClientFactory" />
+  </ItemGroup>
+</Project>

Apis/page-fetcher-api/Controllers/PageController.cs

@@ -0,0 +1,47 @@
+using Api.Services;
+using Microsoft.AspNetCore.Mvc;
+using PageFetcher.Models;
+using Swashbuckle.AspNetCore.Annotations;
+
+namespace Api.Controllers;
+
+/// <summary>
+/// Handles page-fetch requests: navigates to the URL via Playwright and returns rendered HTML and extracted text.
+/// </summary>
+[ApiController]
+[Route("api/page")]
+public sealed class PageController : ControllerBase
+{
+    private readonly PageFetcherService _service;
+    private readonly ILogger<PageController> _logger;
+
+    public PageController(PageFetcherService service, ILogger<PageController> logger)
+    {
+        _service = service;
+        _logger = logger;
+    }
+
+    /// <summary>
+    /// Fetches a web page via headless Chromium.
+    /// Returns rendered HTML and extracted plain text.
+    /// </summary>
+    [HttpPost("fetch")]
+    [SwaggerOperation(Summary = "Fetch a web page", Description = "Navigates to the given URL using Playwright, returns rendered HTML and stripped plain text.")]
+    [SwaggerResponse(StatusCodes.Status200OK, "Page fetched successfully", typeof(FetchPageResponse))]
+    [SwaggerResponse(StatusCodes.Status400BadRequest, "Invalid or non-HTTP(S) URL")]
+    public async Task<ActionResult<FetchPageResponse>> Fetch([FromBody] FetchPageRequest request, CancellationToken ct)
+    {
+        if (string.IsNullOrWhiteSpace(request.Url))
+            return BadRequest(new { Error = "Url is required." });
+
+        if (!Uri.TryCreate(request.Url, UriKind.Absolute, out var uri) ||
+            (uri.Scheme != Uri.UriSchemeHttp && uri.Scheme != Uri.UriSchemeHttps))
+            return BadRequest(new { Error = "Url must be an absolute HTTP or HTTPS URL." });
+
+        _logger.LogInformation("Fetch request: {Url} | caller={Caller} | waitFor={WaitFor}",
+            request.Url, request.CallerService, request.WaitFor);
+
+        var result = await _service.FetchAsync(request, ct);
+        return Ok(result);
+    }
+}

Apis/page-fetcher-api/Dockerfile

@@ -0,0 +1,50 @@
+FROM mcr.microsoft.com/dotnet/sdk:10.0 AS build
+ARG BUILD_CONFIGURATION=Release
+WORKDIR /src
+COPY Directory.Packages.props ./
+
+COPY Apis/page-fetcher-api/page-fetcher-api.csproj Apis/page-fetcher-api/
+COPY Apis/page-fetcher-data/page-fetcher-data.csproj Apis/page-fetcher-data/
+COPY Apis/page-fetcher-api-models/page-fetcher-api-models.csproj Apis/page-fetcher-api-models/
+COPY Apis/common/common.csproj Apis/common/
+COPY Apis/shared-data/shared-data.csproj Apis/shared-data/
+COPY Helpers/startup-helpers/startup-helpers.csproj Helpers/startup-helpers/
+COPY Helpers/common-helpers/common-helpers.csproj Helpers/common-helpers/
+
+RUN dotnet restore Apis/page-fetcher-api/page-fetcher-api.csproj
+
+COPY Apis/page-fetcher-api/ Apis/page-fetcher-api/
+COPY Apis/page-fetcher-data/ Apis/page-fetcher-data/
+COPY Apis/page-fetcher-api-models/ Apis/page-fetcher-api-models/
+COPY Apis/common/ Apis/common/
+COPY Apis/shared-data/ Apis/shared-data/
+COPY Helpers/startup-helpers/ Helpers/startup-helpers/
+COPY Helpers/common-helpers/ Helpers/common-helpers/
+
+RUN dotnet publish Apis/page-fetcher-api/page-fetcher-api.csproj -c $BUILD_CONFIGURATION -o /app/publish /p:UseAppHost=false
+
+# Download Playwright Chromium browser in the build stage.
+# Node.js is only needed here to run npx — it is not copied to the final image.
+ENV PLAYWRIGHT_BROWSERS_PATH=/ms-playwright
+RUN apt-get update && apt-get install -y --no-install-recommends nodejs npm \
+    && npx --yes playwright@1.60.0 install chromium \
+    && rm -rf /var/lib/apt/lists/*
+
+FROM mcr.microsoft.com/dotnet/aspnet:10.0 AS final
+WORKDIR /app
+
+# System libraries required by Chromium on Debian bookworm
+RUN apt-get update && apt-get install -y --no-install-recommends \
+    libnss3 libnspr4 libatk1.0-0 libatk-bridge2.0-0 libcups2 libdrm2 \
+    libxkbcommon0 libxcomposite1 libxdamage1 libxfixes3 libxrandr2 \
+    libgbm1 libasound2t64 libpango-1.0-0 libcairo2 libatspi2.0-0 \
+    libwayland-client0 libx11-xcb1 libx11-6 libxcb1 libxext6 \
+    && rm -rf /var/lib/apt/lists/*
+
+# Copy the Playwright Chromium browser from the build stage
+ENV PLAYWRIGHT_BROWSERS_PATH=/ms-playwright
+COPY --from=build /ms-playwright /ms-playwright
+
+COPY --from=build /app/publish .
+
+ENTRYPOINT ["dotnet", "page-fetcher-api.dll"]

Apis/page-fetcher-api/Program.cs

@@ -0,0 +1,75 @@
+using System.Reflection;
+using Microsoft.EntityFrameworkCore;
+using PageFetcher.Data;
+using Api.Services;
+using PageFetcher.Models.Settings;
+using Serilog;
+using StartupHelpers;
+
+StartupExtensions.LoadDotEnvFile();
+
+const string ServiceName = "page-fetcher-api";
+var appVersion = StartupExtensions.GetApplicationVersion(Assembly.GetExecutingAssembly());
+
+try
+{
+    var builder = WebApplication.CreateBuilder(args);
+
+    builder.ConfigureJsonSerilog(ServiceName, appVersion);
+    Log.Information("Starting {Service} version {AppVersion}", ServiceName, appVersion);
+
+    builder.AddAzureKeyVaultIfConfigured();
+
+    builder.Services.Configure<PageFetcherSettings>(builder.Configuration.GetSection("PageFetcher"));
+
+    builder.Services.AddDbContext<PageFetchDbContext>(options =>
+    {
+        var connectionString = builder.Services.GetConfiguredDbConnectionString(builder.Configuration);
+        options.UseSqlServer(connectionString, sql =>
+        {
+            sql.MigrationsHistoryTable(PageFetchDbContext.MigrationTableName, PageFetchDbContext.SchemaName);
+            sql.MigrationsAssembly("page-fetcher-data");
+        });
+    });
+
+    // Playwright browser: singleton hosted service, shared across all requests
+    builder.Services.AddSingleton<PlaywrightBrowserService>();
+    builder.Services.AddHostedService(sp => sp.GetRequiredService<PlaywrightBrowserService>());
+
+    builder.Services.AddScoped<PageFetcherService>();
+
+    builder.Services.AddControllers();
+    builder.Services.AddSwaggerWithXmlComments(Assembly.GetExecutingAssembly(), "Page Fetcher API");
+
+    var app = builder.Build();
+
+    app.LogStartupDiagnostics(ServiceName);
+
+    app.UseDefaultSerilogRequestLogging();
+    app.UseJsonExceptionHandler(ServiceName);
+    app.UseInternalApiKeyProtection();
+    app.UseSwaggerInDevelopment("Page Fetcher API", "PageFetcherAPI");
+
+    app.UseRouting();
+    app.UseAuthorization();
+    app.MapControllers();
+
+    Log.Information("Running EF Core migrations if any");
+    using (var scope = app.Services.CreateScope())
+    {
+        var db = scope.ServiceProvider.GetRequiredService<PageFetchDbContext>();
+        db.Database.Migrate();
+    }
+
+    Log.Information("{Service} startup complete. Listening for requests...", ServiceName);
+    app.Run();
+}
+catch (Exception ex)
+{
+    Log.Fatal(ex, "{Service} terminated unexpectedly", ServiceName);
+}
+finally
+{
+    Log.Information("Shutting down {Service}", ServiceName);
+    Log.CloseAndFlush();
+}

Apis/page-fetcher-api/Properties/launchSettings.json

@@ -0,0 +1,12 @@
+{
+  "profiles": {
+    "page-fetcher-api": {
+      "commandName": "Project",
+      "launchBrowser": true,
+      "environmentVariables": {
+        "ASPNETCORE_ENVIRONMENT": "Development"
+      },
+      "applicationUrl": "https://localhost:50268;http://localhost:50269"
+    }
+  }
+}

Apis/page-fetcher-api/Services/PageFetcherService.cs

@@ -0,0 +1,144 @@
+using System.Diagnostics;
+using System.Net;
+using System.Text.RegularExpressions;
+using Microsoft.Extensions.Options;
+using Microsoft.Playwright;
+using PageFetcher.Data;
+using PageFetcher.Data.Entities;
+using PageFetcher.Models;
+using PageFetcher.Models.Settings;
+
+namespace Api.Services;
+
+/// <summary>
+/// Fetches a web page via Playwright, extracts plain text, persists the result to the database,
+/// and returns a <see cref="FetchPageResponse"/>.
+/// </summary>
+public sealed class PageFetcherService
+{
+    private readonly PlaywrightBrowserService _browserService;
+    private readonly PageFetchDbContext _db;
+    private readonly PageFetcherSettings _settings;
+    private readonly ILogger<PageFetcherService> _logger;
+
+    public PageFetcherService(
+        PlaywrightBrowserService browserService,
+        PageFetchDbContext db,
+        IOptions<PageFetcherSettings> settings,
+        ILogger<PageFetcherService> logger)
+    {
+        _browserService = browserService;
+        _db = db;
+        _settings = settings.Value;
+        _logger = logger;
+    }
+
+    /// <summary>
+    /// Fetches the page at <paramref name="request.Url"/> using Playwright, saves the fetch record,
+    /// and returns the HTML and extracted text.
+    /// Returns a failed response (with <see cref="FetchPageResponse.Success"/> = false) rather than throwing
+    /// on network or navigation errors.
+    /// </summary>
+    public async Task<FetchPageResponse> FetchAsync(FetchPageRequest request, CancellationToken ct)
+    {
+        var sw = Stopwatch.StartNew();
+        string html = string.Empty;
+        string text = string.Empty;
+        int? statusCode = null;
+        bool success = false;
+        string? errorMessage = null;
+        string finalUrl = request.Url;
+
+        try
+        {
+            var page = await _browserService.Browser.NewPageAsync();
+            await using var _ = page.ConfigureAwait(false);
+
+            var waitUntil = request.WaitFor?.ToLowerInvariant() switch
+            {
+                "load" => WaitUntilState.Load,
+                "domcontentloaded" => WaitUntilState.DOMContentLoaded,
+                _ => WaitUntilState.NetworkIdle
+            };
+
+            IResponse? response;
+            try
+            {
+                response = await page.GotoAsync(request.Url, new PageGotoOptions
+                {
+                    WaitUntil = waitUntil,
+                    Timeout = _settings.TimeoutSeconds * 1_000
+                });
+            }
+            catch (TimeoutException)
+            {
+                _logger.LogWarning("Playwright NetworkIdle timeout for {Url}, using partial content", request.Url);
+                response = null;
+            }
+
+            statusCode = response?.Status;
+            finalUrl = page.Url;
+            html = await page.ContentAsync();
+            text = ExtractText(html);
+            success = true;
+
+            _logger.LogInformation("Fetched {Url} → HTTP {Status} | HTML {HtmlLen} chars | text {TextLen} chars | {DurationMs} ms",
+                request.Url, statusCode?.ToString() ?? "timeout", html.Length, text.Length, sw.ElapsedMilliseconds);
+        }
+        catch (Exception ex)
+        {
+            errorMessage = ex.Message;
+            _logger.LogError(ex, "Failed to fetch {Url}", request.Url);
+        }
+        finally
+        {
+            sw.Stop();
+        }
+
+        // Persist fetch record
+        var entity = new PageFetchEntity
+        {
+            Id = Guid.NewGuid().ToString("N"),
+            Url = request.Url,
+            CallerService = request.CallerService ?? string.Empty,
+            HttpStatusCode = statusCode,
+            Html = html,
+            Text = text,
+            DurationMs = sw.ElapsedMilliseconds,
+            Success = success,
+            ErrorMessage = errorMessage
+        };
+
+        _db.PageFetches.Add(entity);
+        await _db.SaveChangesAsync(ct);
+
+        return new FetchPageResponse
+        {
+            Url = finalUrl,
+            StatusCode = statusCode ?? 0,
+            Html = html,
+            Text = text,
+            Success = success,
+            Error = errorMessage
+        };
+    }
+
+    /// <summary>
+    /// Strips script/style blocks and all HTML tags from raw HTML, normalises whitespace,
+    /// and truncates to <see cref="PageFetcherSettings.MaxTextChars"/>.
+    /// </summary>
+    private string ExtractText(string html)
+    {
+        if (string.IsNullOrWhiteSpace(html)) return string.Empty;
+
+        var text = html;
+        text = Regex.Replace(text, "<script[\\s\\S]*?</script>", " ", RegexOptions.IgnoreCase);
+        text = Regex.Replace(text, "<style[\\s\\S]*?</style>", " ", RegexOptions.IgnoreCase);
+        text = Regex.Replace(text, "<[^>]+>", " ");
+        text = WebUtility.HtmlDecode(text);
+        text = string.Join(' ', text.Split((char[]?)null, StringSplitOptions.RemoveEmptyEntries)).Trim();
+
+        var max = Math.Max(4_000, _settings.MaxTextChars);
+        return text.Length <= max ? text : text[..max];
+    }
+}

Apis/page-fetcher-api/Services/PlaywrightBrowserService.cs

@@ -0,0 +1,49 @@
+using Microsoft.Playwright;
+
+namespace Api.Services;
+
+/// <summary>
+/// Singleton hosted service that owns the Playwright Chromium browser process for the lifetime of the application.
+/// Launches the browser once at startup and exposes it for injection into <see cref="PageFetcherService"/>.
+/// </summary>
+public sealed class PlaywrightBrowserService : IHostedService, IAsyncDisposable
+{
+    private IPlaywright? _playwright;
+    private IBrowser? _browser;
+    private readonly ILogger<PlaywrightBrowserService> _logger;
+
+    public PlaywrightBrowserService(ILogger<PlaywrightBrowserService> logger)
+    {
+        _logger = logger;
+    }
+
+    /// <summary>The running Chromium browser instance. Available after <see cref="StartAsync"/> completes.</summary>
+    public IBrowser Browser => _browser ?? throw new InvalidOperationException("Browser has not been started yet.");
+
+    /// <inheritdoc />
+    public async Task StartAsync(CancellationToken cancellationToken)
+    {
+        _logger.LogInformation("Launching Playwright Chromium browser...");
+        _playwright = await Playwright.CreateAsync();
+        _browser = await _playwright.Chromium.LaunchAsync(new BrowserTypeLaunchOptions
+        {
+            Headless = true,
+            Args = ["--no-sandbox", "--disable-setuid-sandbox", "--disable-dev-shm-usage"]
+        });
+        _logger.LogInformation("Playwright Chromium browser launched successfully.");
+    }
+
+    /// <inheritdoc />
+    public async Task StopAsync(CancellationToken cancellationToken)
+    {
+        _logger.LogInformation("Closing Playwright Chromium browser...");
+        if (_browser is not null) await _browser.CloseAsync();
+    }
+
+    /// <inheritdoc />
+    public async ValueTask DisposeAsync()
+    {
+        if (_browser is not null) await _browser.DisposeAsync();
+        _playwright?.Dispose();
+    }
+}

Apis/page-fetcher-api/appsettings.json

@@ -0,0 +1,73 @@
+{
+  "Serilog": {
+    "Using": [
+      "Serilog.Sinks.Console",
+      "Serilog.Sinks.File"
+    ],
+    "MinimumLevel": {
+      "Default": "Information",
+      "Override": {
+        "Microsoft.AspNetCore": "Warning",
+        "Microsoft.AspNetCore.Hosting": "Information",
+        "Microsoft.AspNetCore.Routing": "Warning",
+        "System.Net.Http.HttpClient": "Warning",
+        "PageFetcherApi": "Information"
+      }
+    },
+    "WriteTo": [
+      {
+        "Name": "Console",
+        "Args": {
+          "outputTemplate": "[{Timestamp:HH:mm:ss} {Level:u3}] {SourceContext}: {Message:lj}{NewLine}{Exception}"
+        }
+      },
+      {
+        "Name": "File",
+        "Args": {
+          "path": "logs/page-fetcher-api-.log",
+          "rollingInterval": "Day",
+          "retainedFileCountLimit": 30,
+          "outputTemplate": "{Timestamp:yyyy-MM-dd HH:mm:ss.fff zzz} [{Level:u3}] {SourceContext}: {Message:lj}{NewLine}{Exception}"
+        }
+      }
+    ],
+    "Enrich": [
+      "FromLogContext",
+      "WithMachineName",
+      "WithEnvironmentName"
+    ]
+  },
+  "Logging": {
+    "LogLevel": {
+      "Default": "Information",
+      "Microsoft.AspNetCore": "Warning",
+      "Microsoft.AspNetCore.Hosting": "Information",
+      "Microsoft.AspNetCore.Routing": "Warning",
+      "System.Net.Http.HttpClient": "Warning",
+      "PageFetcherApi": "Information"
+    }
+  },
+  "LogEnvironmentOnStartup": true,
+  "AllowedHosts": "*",
+  "KeyVault": {
+    "VaultUri": "",
+    "Enabled": false
+  },
+  "Database": {
+    "Host": "localhost",
+    "Port": 1433,
+    "Name": "MyAiDb",
+    "User": "sa",
+    "Password": "",
+    "TrustServerCertificate": true
+  },
+  "InternalApi": {
+    "ApiKey": "",
+    "RequireApiKey": true
+  },
+  "PageFetcher": {
+    "DefaultWaitFor": "networkidle",
+    "TimeoutSeconds": 30,
+    "MaxTextChars": 60000
+  }
+}

Apis/page-fetcher-api/page-fetcher-api.csproj

@@ -0,0 +1,34 @@
+<Project Sdk="Microsoft.NET.Sdk.Web">
+  <PropertyGroup>
+    <TargetFramework>net10.0</TargetFramework>
+    <Nullable>enable</Nullable>
+    <ImplicitUsings>enable</ImplicitUsings>
+    <DockerDefaultTargetOS>Linux</DockerDefaultTargetOS>
+    <RootNamespace>PageFetcherApi</RootNamespace>
+    <GenerateDocumentationFile>true</GenerateDocumentationFile>
+    <NoWarn>$(NoWarn);1591</NoWarn>
+  </PropertyGroup>
+  <ItemGroup>
+    <PackageReference Include="Microsoft.Playwright" />
+    <PackageReference Include="Microsoft.EntityFrameworkCore.SqlServer" />
+    <PackageReference Include="Microsoft.EntityFrameworkCore.Design">
+      <PrivateAssets>all</PrivateAssets>
+      <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
+    </PackageReference>
+    <PackageReference Include="Serilog.AspNetCore" />
+    <PackageReference Include="Serilog.Enrichers.Environment" />
+    <PackageReference Include="Serilog.Sinks.Console" />
+    <PackageReference Include="Serilog.Sinks.File" />
+    <PackageReference Include="Serilog.Sinks.Email" />
+    <PackageReference Include="Swashbuckle.AspNetCore" />
+    <PackageReference Include="Swashbuckle.AspNetCore.Annotations" />
+    <PackageReference Include="DotNetEnv" />
+  </ItemGroup>
+  <ItemGroup>
+    <ProjectReference Include="..\page-fetcher-data\page-fetcher-data.csproj" />
+    <ProjectReference Include="..\page-fetcher-api-models\page-fetcher-api-models.csproj" />
+    <ProjectReference Include="..\common\common.csproj" />
+    <ProjectReference Include="..\..\Helpers\startup-helpers\startup-helpers.csproj" />
+    <ProjectReference Include="..\..\Helpers\common-helpers\common-helpers.csproj" />
+  </ItemGroup>
+</Project>

Apis/page-fetcher-data/Data/Entities/PageFetchEntity.cs

@@ -0,0 +1,34 @@
+using Shared.Data.Entities;
+
+namespace PageFetcher.Data.Entities;
+
+/// <summary>
+/// Audit record of a single page-fetch operation performed by the page-fetcher-api.
+/// Stores the full rendered HTML and extracted plain text for every URL fetched.
+/// </summary>
+public sealed class PageFetchEntity : BaseEntity
+{
+    /// <summary>The URL that was requested.</summary>
+    public string Url { get; set; } = string.Empty;
+
+    /// <summary>Name of the service that requested the fetch (e.g. <c>cv-matcher-api</c>, <c>cv-search-job</c>).</summary>
+    public string CallerService { get; set; } = string.Empty;
+
+    /// <summary>HTTP status code returned by the remote server. <c>null</c> on network failure.</summary>
+    public int? HttpStatusCode { get; set; }
+
+    /// <summary>Full rendered HTML as returned by Playwright.</summary>
+    public string Html { get; set; } = string.Empty;
+
+    /// <summary>Plain text extracted from the HTML (script/style stripped, whitespace normalised).</summary>
+    public string Text { get; set; } = string.Empty;
+
+    /// <summary>Playwright round-trip time in milliseconds.</summary>
+    public long DurationMs { get; set; }
+
+    /// <summary><c>true</c> when the page was fetched successfully; <c>false</c> on timeout or network error.</summary>
+    public bool Success { get; set; }
+
+    /// <summary>Exception message when <see cref="Success"/> is <c>false</c>.</summary>
+    public string? ErrorMessage { get; set; }
+}

Apis/page-fetcher-data/MigrationConstants.cs

@@ -0,0 +1,8 @@
+namespace PageFetcher.Data;
+
+/// <summary>Schema and migration-history table name constants for the pageFetcher EF schema.</summary>
+public static class MigrationConstants
+{
+    public const string SchemaName = "pageFetcher";
+    public const string MigrationTableName = "_Migrations";
+}

Apis/page-fetcher-data/Migrations/20260608143523_InitialSchema.Designer.cs

@@ -0,0 +1,82 @@
+// <auto-generated />
+using System;
+using Microsoft.EntityFrameworkCore;
+using Microsoft.EntityFrameworkCore.Infrastructure;
+using Microsoft.EntityFrameworkCore.Metadata;
+using Microsoft.EntityFrameworkCore.Migrations;
+using Microsoft.EntityFrameworkCore.Storage.ValueConversion;
+using PageFetcher.Data;
+
+#nullable disable
+
+namespace PageFetcher.Data.Migrations
+{
+    [DbContext(typeof(PageFetchDbContext))]
+    [Migration("20260608143523_InitialSchema")]
+    partial class InitialSchema
+    {
+        /// <inheritdoc />
+        protected override void BuildTargetModel(ModelBuilder modelBuilder)
+        {
+#pragma warning disable 612, 618
+            modelBuilder
+                .HasDefaultSchema("pageFetcher")
+                .HasAnnotation("ProductVersion", "10.0.7")
+                .HasAnnotation("Relational:MaxIdentifierLength", 128);
+
+            SqlServerModelBuilderExtensions.UseIdentityColumns(modelBuilder);
+
+            modelBuilder.Entity("PageFetcher.Data.Entities.PageFetchEntity", b =>
+                {
+                    b.Property<string>("Id")
+                        .HasMaxLength(64)
+                        .HasColumnType("nvarchar(64)");
+
+                    b.Property<string>("CallerService")
+                        .IsRequired()
+                        .HasMaxLength(64)
+                        .HasColumnType("nvarchar(64)");
+
+                    b.Property<DateTime>("CreatedAt")
+                        .ValueGeneratedOnAdd()
+                        .HasColumnType("datetime2")
+                        .HasDefaultValueSql("SYSUTCDATETIME()");
+
+                    b.Property<long>("DurationMs")
+                        .HasColumnType("bigint");
+
+                    b.Property<string>("ErrorMessage")
+                        .HasMaxLength(2000)
+                        .HasColumnType("nvarchar(2000)");
+
+                    b.Property<string>("Html")
+                        .IsRequired()
+                        .HasColumnType("nvarchar(max)");
+
+                    b.Property<int?>("HttpStatusCode")
+                        .HasColumnType("int");
+
+                    b.Property<bool>("Success")
+                        .HasColumnType("bit");
+
+                    b.Property<string>("Text")
+                        .IsRequired()
+                        .HasColumnType("nvarchar(max)");
+
+                    b.Property<string>("Url")
+                        .IsRequired()
+                        .HasMaxLength(2000)
+                        .HasColumnType("nvarchar(2000)");
+
+                    b.HasKey("Id");
+
+                    b.HasIndex("CreatedAt");
+
+                    b.HasIndex("Url");
+
+                    b.ToTable("PageFetches", "pageFetcher");
+                });
+#pragma warning restore 612, 618
+        }
+    }
+}

Apis/page-fetcher-data/Migrations/20260608143523_InitialSchema.cs

@@ -0,0 +1,59 @@
+using System;
+using Microsoft.EntityFrameworkCore.Migrations;
+
+#nullable disable
+
+namespace PageFetcher.Data.Migrations
+{
+    /// <inheritdoc />
+    public partial class InitialSchema : Migration
+    {
+        /// <inheritdoc />
+        protected override void Up(MigrationBuilder migrationBuilder)
+        {
+            migrationBuilder.EnsureSchema(
+                name: MigrationConstants.SchemaName);
+
+            migrationBuilder.CreateTable(
+                name: "PageFetches",
+                schema: MigrationConstants.SchemaName,
+                columns: table => new
+                {
+                    Id = table.Column<string>(type: "nvarchar(64)", maxLength: 64, nullable: false),
+                    Url = table.Column<string>(type: "nvarchar(2000)", maxLength: 2000, nullable: false),
+                    CallerService = table.Column<string>(type: "nvarchar(64)", maxLength: 64, nullable: false),
+                    HttpStatusCode = table.Column<int>(type: "int", nullable: true),
+                    Html = table.Column<string>(type: "nvarchar(max)", nullable: false),
+                    Text = table.Column<string>(type: "nvarchar(max)", nullable: false),
+                    DurationMs = table.Column<long>(type: "bigint", nullable: false),
+                    Success = table.Column<bool>(type: "bit", nullable: false),
+                    ErrorMessage = table.Column<string>(type: "nvarchar(2000)", maxLength: 2000, nullable: true),
+                    CreatedAt = table.Column<DateTime>(type: "datetime2", nullable: false, defaultValueSql: "SYSUTCDATETIME()")
+                },
+                constraints: table =>
+                {
+                    table.PrimaryKey("PK_PageFetches", x => x.Id);
+                });
+
+            migrationBuilder.CreateIndex(
+                name: "IX_PageFetches_CreatedAt",
+                schema: MigrationConstants.SchemaName,
+                table: "PageFetches",
+                column: "CreatedAt");
+
+            migrationBuilder.CreateIndex(
+                name: "IX_PageFetches_Url",
+                schema: MigrationConstants.SchemaName,
+                table: "PageFetches",
+                column: "Url");
+        }
+
+        /// <inheritdoc />
+        protected override void Down(MigrationBuilder migrationBuilder)
+        {
+            migrationBuilder.DropTable(
+                name: "PageFetches",
+                schema: MigrationConstants.SchemaName);
+        }
+    }
+}

Apis/page-fetcher-data/Migrations/PageFetchDbContextModelSnapshot.cs

@@ -0,0 +1,79 @@
+// <auto-generated />
+using System;
+using Microsoft.EntityFrameworkCore;
+using Microsoft.EntityFrameworkCore.Infrastructure;
+using Microsoft.EntityFrameworkCore.Metadata;
+using Microsoft.EntityFrameworkCore.Storage.ValueConversion;
+using PageFetcher.Data;
+
+#nullable disable
+
+namespace PageFetcher.Data.Migrations
+{
+    [DbContext(typeof(PageFetchDbContext))]
+    partial class PageFetchDbContextModelSnapshot : ModelSnapshot
+    {
+        protected override void BuildModel(ModelBuilder modelBuilder)
+        {
+#pragma warning disable 612, 618
+            modelBuilder
+                .HasDefaultSchema("pageFetcher")
+                .HasAnnotation("ProductVersion", "10.0.7")
+                .HasAnnotation("Relational:MaxIdentifierLength", 128);
+
+            SqlServerModelBuilderExtensions.UseIdentityColumns(modelBuilder);
+
+            modelBuilder.Entity("PageFetcher.Data.Entities.PageFetchEntity", b =>
+                {
+                    b.Property<string>("Id")
+                        .HasMaxLength(64)
+                        .HasColumnType("nvarchar(64)");
+
+                    b.Property<string>("CallerService")
+                        .IsRequired()
+                        .HasMaxLength(64)
+                        .HasColumnType("nvarchar(64)");
+
+                    b.Property<DateTime>("CreatedAt")
+                        .ValueGeneratedOnAdd()
+                        .HasColumnType("datetime2")
+                        .HasDefaultValueSql("SYSUTCDATETIME()");
+
+                    b.Property<long>("DurationMs")
+                        .HasColumnType("bigint");
+
+                    b.Property<string>("ErrorMessage")
+                        .HasMaxLength(2000)
+                        .HasColumnType("nvarchar(2000)");
+
+                    b.Property<string>("Html")
+                        .IsRequired()
+                        .HasColumnType("nvarchar(max)");
+
+                    b.Property<int?>("HttpStatusCode")
+                        .HasColumnType("int");
+
+                    b.Property<bool>("Success")
+                        .HasColumnType("bit");
+
+                    b.Property<string>("Text")
+                        .IsRequired()
+                        .HasColumnType("nvarchar(max)");
+
+                    b.Property<string>("Url")
+                        .IsRequired()
+                        .HasMaxLength(2000)
+                        .HasColumnType("nvarchar(2000)");
+
+                    b.HasKey("Id");
+
+                    b.HasIndex("CreatedAt");
+
+                    b.HasIndex("Url");
+
+                    b.ToTable("PageFetches", "pageFetcher");
+                });
+#pragma warning restore 612, 618
+        }
+    }
+}

Apis/page-fetcher-data/PageFetchDbContext.cs

@@ -0,0 +1,45 @@
+using Microsoft.EntityFrameworkCore;
+using PageFetcher.Data.Entities;
+
+namespace PageFetcher.Data;
+
+/// <summary>
+/// EF Core DbContext for the <c>pageFetcher</c> schema.
+/// Owns the <c>PageFetches</c> audit table.
+/// </summary>
+public sealed class PageFetchDbContext : DbContext
+{
+    public const string SchemaName = MigrationConstants.SchemaName;
+    public const string MigrationTableName = MigrationConstants.MigrationTableName;
+
+    public PageFetchDbContext(DbContextOptions<PageFetchDbContext> options) : base(options) { }
+
+    public DbSet<PageFetchEntity> PageFetches => Set<PageFetchEntity>();
+
+    protected override void OnConfiguring(DbContextOptionsBuilder optionsBuilder)
+    {
+        base.OnConfiguring(optionsBuilder);
+        optionsBuilder.UseSqlServer(x => x.MigrationsHistoryTable(MigrationTableName, SchemaName));
+    }
+
+    protected override void OnModelCreating(ModelBuilder modelBuilder)
+    {
+        modelBuilder.HasDefaultSchema(SchemaName);
+
+        modelBuilder.Entity<PageFetchEntity>(entity =>
+        {
+            entity.ToTable("PageFetches");
+            entity.HasKey(x => x.Id);
+            entity.Property(x => x.Id).HasMaxLength(64);
+            entity.Property(x => x.Url).HasMaxLength(2000).IsRequired();
+            entity.Property(x => x.CallerService).HasMaxLength(64).IsRequired();
+            entity.Property(x => x.Html).IsRequired();
+            entity.Property(x => x.Text).IsRequired();
+            entity.Property(x => x.ErrorMessage).HasMaxLength(2000);
+            entity.Property(x => x.CreatedAt).HasDefaultValueSql("SYSUTCDATETIME()");
+
+            entity.HasIndex(x => x.Url);
+            entity.HasIndex(x => x.CreatedAt);
+        });
+    }
+}

Apis/page-fetcher-data/page-fetcher-data.csproj

@@ -0,0 +1,19 @@
+<Project Sdk="Microsoft.NET.Sdk">
+  <PropertyGroup>
+    <TargetFramework>net10.0</TargetFramework>
+    <Nullable>enable</Nullable>
+    <ImplicitUsings>enable</ImplicitUsings>
+    <AssemblyName>page-fetcher-data</AssemblyName>
+    <RootNamespace>PageFetcher.Data</RootNamespace>
+  </PropertyGroup>
+  <ItemGroup>
+    <PackageReference Include="Microsoft.EntityFrameworkCore.SqlServer" />
+    <PackageReference Include="Microsoft.EntityFrameworkCore.Design">
+      <PrivateAssets>all</PrivateAssets>
+      <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
+    </PackageReference>
+  </ItemGroup>
+  <ItemGroup>
+    <ProjectReference Include="..\shared-data\shared-data.csproj" />
+  </ItemGroup>
+</Project>

Directory.Packages.props

@@ -18,7 +18,7 @@
     <!-- Config -->
     <PackageVersion Include="DotNetEnv" Version="3.2.0" />
     <!-- HTTP / Refit -->
-    <PackageVersion Include="Refit.HttpClientFactory" Version="10.1.6" />
+    <PackageVersion Include="Refit.HttpClientFactory" Version="11.0.1" />
     <!-- Serilog -->
     <PackageVersion Include="Serilog.AspNetCore" Version="10.0.0" />
     <PackageVersion Include="Serilog.Enrichers.Environment" Version="3.0.1" />

Jobs/cv-search-job/Dockerfile

@@ -9,6 +9,7 @@ COPY Apis/cv-search-data/cv-search-data.csproj Apis/cv-search-data/
 COPY Apis/cv-matcher-api-models/cv-matcher-api-models.csproj Apis/cv-matcher-api-models/
 COPY Apis/email-data/email-data.csproj Apis/email-data/
 COPY Apis/email-api-models/email-api-models.csproj Apis/email-api-models/
+COPY Apis/page-fetcher-api-models/page-fetcher-api-models.csproj Apis/page-fetcher-api-models/
 COPY Apis/common/common.csproj Apis/common/
 COPY Apis/myai-data/myai-data.csproj Apis/myai-data/
 COPY Apis/shared-data/shared-data.csproj Apis/shared-data/
@@ -22,6 +23,7 @@ COPY Apis/cv-search-data/ Apis/cv-search-data/
 COPY Apis/cv-matcher-api-models/ Apis/cv-matcher-api-models/
 COPY Apis/email-data/ Apis/email-data/
 COPY Apis/email-api-models/ Apis/email-api-models/
+COPY Apis/page-fetcher-api-models/ Apis/page-fetcher-api-models/
 COPY Apis/common/ Apis/common/
 COPY Apis/myai-data/ Apis/myai-data/
 COPY Apis/shared-data/ Apis/shared-data/

Jobs/cv-search-job/Program.cs

@@ -7,13 +7,14 @@ using Email.Data;
 using Email.Data.Repositories;
 using Email.Data.Repositories.Contracts;
 using Email.Data.Services;
-using EmailApi.Models.Clients;
+using Email.Models.Clients;
 using CvSearchJob.Tasks;
 using JobScheduler.Scheduling;
 using JobScheduler.Tasks;
 using Microsoft.EntityFrameworkCore;
 using Microsoft.Extensions.DependencyInjection;
 using Microsoft.Extensions.Hosting;
+using PageFetcher.Models;
 using Refit;
 using Serilog;
 using Common.Settings;
@@ -81,7 +82,19 @@ try
                 client.DefaultRequestHeaders.Add("X-Internal-Api-Key", key);
         });
 
-    builder.Services.AddHttpClient<HtmlJobSearcher>();
+    builder.Services.AddRefitClient<IPageFetcherApiClient>()
+        .ConfigureHttpClient((sp, client) =>
+        {
+            var config = sp.GetRequiredService<Microsoft.Extensions.Configuration.IConfiguration>();
+            var baseUrl = config["PageFetcherApi:BaseUrl"] ?? string.Empty;
+            if (!string.IsNullOrWhiteSpace(baseUrl))
+                client.BaseAddress = new Uri(baseUrl.TrimEnd('/') + "/");
+            var key = config["PageFetcherApi:InternalApiKey"];
+            if (!string.IsNullOrWhiteSpace(key))
+                client.DefaultRequestHeaders.Add("X-Internal-Api-Key", key);
+        });
+
+    builder.Services.AddSingleton<HtmlJobSearcher>();
     builder.Services.AddSingleton<CvSearchEmailSender>();
 
     builder.Services.AddSingleton<CvSearchJobTask>();

Jobs/cv-search-job/Services/CvSearchEmailSender.cs

@@ -1,8 +1,8 @@
 using CvMatcher.Models.Responses;
 using CvSearch.Data.Entities;
 using Email.Data.Services;
-using EmailApi.Models.Clients;
-using EmailApi.Models.Requests;
+using Email.Models.Clients;
+using Email.Models.Requests;
 using Microsoft.Extensions.Logging;
 
 namespace CvSearchJob.Services;

Jobs/cv-search-job/Services/HtmlJobSearcher.cs

@@ -1,36 +1,39 @@
 using System.Text.RegularExpressions;
 using System.Web;
 using CvMatcher.Models.Settings;
-using Microsoft.Playwright;
+using PageFetcher.Models;
 using Microsoft.Extensions.Logging;
 
 namespace CvSearchJob.Services;
 
 /// <summary>
-/// Config-driven HTML scraper that fetches a provider's job listing page and extracts matching job URLs.
-/// Uses a two-stage anchor filter: href must contain the provider's link pattern, and anchor text must
-/// contain at least one CV keyword.
-/// Supports both plain HTTP GET (default) and headless Chromium rendering for JS-heavy SPAs.
+/// A URL and its anchor text as scraped from a job listing search-results page.
+/// </summary>
+public sealed record JobCandidate(string Url, string Title);
+
+/// <summary>
+/// Config-driven HTML scraper that fetches a provider's job listing page via <c>page-fetcher-api</c>
+/// and extracts matching job URL candidates.
+/// Uses a two-stage anchor filter: href must contain the provider's link pattern, and (optionally)
+/// anchor text must contain at least one CV keyword.
 /// </summary>
 public sealed class HtmlJobSearcher
 {
-    private readonly HttpClient _http;
+    private readonly IPageFetcherApiClient _pageFetcher;
     private readonly ILogger<HtmlJobSearcher> _logger;
 
-    public HtmlJobSearcher(HttpClient http, ILogger<HtmlJobSearcher> logger)
+    public HtmlJobSearcher(IPageFetcherApiClient pageFetcher, ILogger<HtmlJobSearcher> logger)
     {
-        _http = http;
+        _pageFetcher = pageFetcher;
         _logger = logger;
-        _http.Timeout = TimeSpan.FromSeconds(20);
-        _http.DefaultRequestHeaders.UserAgent.ParseAdd("Mozilla/5.0 (compatible; MyAi.ro CV-Search/1.0)");
     }
 
     /// <summary>
-    /// Fetches the provider's search result page for the combined initial + CV keywords, parses all anchor
-    /// tags, applies the two-stage filter, and returns up to <see cref="JobProviderConfig.MaxResults"/> absolute URLs.
-    /// Returns an empty list when the HTTP request fails rather than throwing.
+    /// Fetches the provider's search result page, parses all anchor tags, applies the two-stage filter,
+    /// and returns up to <see cref="JobProviderConfig.MaxResults"/> candidates (URL + title).
+    /// Returns an empty list when the page fetch fails rather than throwing.
     /// </summary>
-    public async Task<IReadOnlyList<string>> SearchJobUrlsAsync(
+    public async Task<IReadOnlyList<JobCandidate>> SearchJobUrlsAsync(
         JobProviderConfig provider,
         IReadOnlyList<string> cvKeywords,
         string? location,
@@ -61,24 +64,29 @@ public sealed class HtmlJobSearcher
             .Replace("{location-slug}", locationSlug);
 
         _logger.LogInformation(
-            "Provider {Provider}: fetching {Url} [{Mode}] | CV keywords: [{Keywords}] | Location: {Location}",
+            "Provider {Provider}: fetching {Url} | CV keywords: [{Keywords}] | Location: {Location}",
             provider.Name, searchUrl,
-            provider.UseHeadlessBrowser ? "headless" : "http",
             string.Join(", ", cvKeywords),
             location ?? "(none)");
 
-        string? html;
-        if (provider.UseHeadlessBrowser)
-            html = await FetchWithPlaywrightAsync(provider.Name, searchUrl, ct);
-        else
-            html = await FetchWithHttpAsync(provider.Name, searchUrl, ct);
+        var fetchResponse = await _pageFetcher.FetchAsync(new FetchPageRequest
+        {
+            Url = searchUrl,
+            WaitFor = provider.UseHeadlessBrowser ? "networkidle" : "domcontentloaded",
+            CallerService = "cv-search-job"
+        }, ct);
 
-        if (html is null) return [];
+        if (!fetchResponse.Success || string.IsNullOrWhiteSpace(fetchResponse.Html))
+        {
+            _logger.LogWarning("Provider {Provider}: page fetch failed — {Error}", provider.Name, fetchResponse.Error);
+            return [];
+        }
 
+        var html = fetchResponse.Html;
         _logger.LogInformation("Provider {Provider}: received {Length} chars of HTML", provider.Name, html.Length);
 
         var baseUri = new Uri(searchUrl);
-        var results = new List<string>();
+        var results = new List<JobCandidate>();
         var seen = new HashSet<string>(StringComparer.OrdinalIgnoreCase);
 
         var anchorPattern = new Regex(@"<a[^>]+href=[""']([^""']+)[""'][^>]*>(.*?)</a>",
@@ -123,7 +131,7 @@ public sealed class HtmlJobSearcher
 
             var url = absoluteUri.GetLeftPart(UriPartial.Path);
             if (seen.Add(url))
-                results.Add(url);
+                results.Add(new JobCandidate(url, anchorText));
         }
 
         _logger.LogInformation(
@@ -132,61 +140,4 @@ public sealed class HtmlJobSearcher
 
         return results;
     }
-
-    private async Task<string?> FetchWithHttpAsync(string providerName, string url, CancellationToken ct)
-    {
-        try
-        {
-            return await _http.GetStringAsync(url, ct);
-        }
-        catch (Exception ex)
-        {
-            _logger.LogError(ex, "Provider {Provider}: HTTP fetch failed for {Url}", providerName, url);
-            return null;
-        }
-    }
-
-    private async Task<string?> FetchWithPlaywrightAsync(string providerName, string url, CancellationToken ct)
-    {
-        try
-        {
-            using var playwright = await Playwright.CreateAsync();
-            await using var browser = await playwright.Chromium.LaunchAsync(new BrowserTypeLaunchOptions
-            {
-                Headless = true,
-                Args = ["--no-sandbox", "--disable-setuid-sandbox", "--disable-dev-shm-usage"]
-            });
-
-            var page = await browser.NewPageAsync();
-
-            IResponse? response;
-            try
-            {
-                response = await page.GotoAsync(url, new PageGotoOptions
-                {
-                    WaitUntil = WaitUntilState.NetworkIdle,
-                    Timeout = 30_000
-                });
-            }
-            catch (TimeoutException)
-            {
-                // NetworkIdle timed out — use whatever content rendered so far
-                _logger.LogWarning("Provider {Provider}: Playwright NetworkIdle timeout for {Url}, using partial content", providerName, url);
-                return await page.ContentAsync();
-            }
-
-            if (response is null || response.Status >= 400)
-            {
-                _logger.LogWarning("Provider {Provider}: Playwright got HTTP {Status} for {Url}", providerName, response?.Status, url);
-                return null;
-            }
-
-            return await page.ContentAsync();
-        }
-        catch (Exception ex)
-        {
-            _logger.LogError(ex, "Provider {Provider}: Playwright fetch failed for {Url}", providerName, url);
-            return null;
-        }
-    }
 }

Jobs/cv-search-job/Tasks/CvSearchJobTask.cs

@@ -11,6 +11,7 @@ using Microsoft.Extensions.Configuration;
 using Microsoft.Extensions.DependencyInjection;
 using Microsoft.Extensions.Logging;
 using Microsoft.Extensions.Options;
+using PageFetcher.Models;
 
 namespace CvSearchJob.Tasks;
 
@@ -24,6 +25,7 @@ public sealed class CvSearchJobTask : IJobTask
     private readonly JobSearchSettings _settings;
     private readonly HtmlJobSearcher _searcher;
     private readonly ICvMatcherInternalApi _matcherApi;
+    private readonly IPageFetcherApiClient _pageFetcher;
     private readonly CvSearchEmailSender _emailSender;
     private readonly ILogger<CvSearchJobTask> _logger;
 
@@ -34,6 +36,7 @@ public sealed class CvSearchJobTask : IJobTask
         IOptions<JobSearchSettings> settings,
         HtmlJobSearcher searcher,
         ICvMatcherInternalApi matcherApi,
+        IPageFetcherApiClient pageFetcher,
         CvSearchEmailSender emailSender,
         ILogger<CvSearchJobTask> logger)
     {
@@ -41,6 +44,7 @@ public sealed class CvSearchJobTask : IJobTask
         _settings = settings.Value;
         _searcher = searcher;
         _matcherApi = matcherApi;
+        _pageFetcher = pageFetcher;
         _emailSender = emailSender;
         _logger = logger;
     }
@@ -126,7 +130,8 @@ public sealed class CvSearchJobTask : IJobTask
 
     /// <summary>
     /// Runs the full search pipeline for a session: scrapes all providers, deduplicates URLs,
-    /// scores each candidate via the matcher API, and persists results that meet the minimum score threshold.
+    /// fetches each individual job page via page-fetcher-api, applies a keyword pre-filter,
+    /// scores passing candidates via the matcher API, and persists results that meet the minimum score threshold.
     /// </summary>
     private async Task<List<JobSearchResultEntity>> RunSearchAsync(
         JobSearchSessionEntity session,
@@ -138,30 +143,59 @@ public sealed class CvSearchJobTask : IJobTask
         if (cvKeywords.Count == 0)
             _logger.LogWarning("Session {SessionId}: keyword list is empty — scraper will rely on provider InitialKeywords only", session.Id);
 
-        var jobUrls = new HashSet<string>(StringComparer.OrdinalIgnoreCase);
+        var jobCandidates = new Dictionary<string, string>(StringComparer.OrdinalIgnoreCase); // url → title
 
         foreach (var provider in providers)
         {
-            var urls = await _searcher.SearchJobUrlsAsync(provider, cvKeywords, session.Location, ct);
-            _logger.LogInformation("Session {SessionId}: provider {Provider} returned {Count} URLs", session.Id, provider.Name, urls.Count);
-            foreach (var url in urls) jobUrls.Add(url);
+            var candidates = await _searcher.SearchJobUrlsAsync(provider, cvKeywords, session.Location, ct);
+            _logger.LogInformation("Session {SessionId}: provider {Provider} returned {Count} candidates", session.Id, provider.Name, candidates.Count);
+            foreach (var c in candidates)
+                jobCandidates.TryAdd(c.Url, c.Title);
         }
 
-        var candidates = jobUrls.Take(_settings.MaxJobsToMatch).ToList();
+        var deduped = jobCandidates.Take(_settings.MaxJobsToMatch).ToList();
         _logger.LogInformation(
-            "Session {SessionId}: {Total} unique URLs across all providers, scoring {Scoring} (cap={Cap})",
-            session.Id, jobUrls.Count, candidates.Count, _settings.MaxJobsToMatch);
+            "Session {SessionId}: {Total} unique URLs across all providers, processing up to {Cap}",
+            session.Id, jobCandidates.Count, deduped.Count);
 
         var results = new List<JobSearchResultEntity>();
 
-        foreach (var url in candidates)
+        foreach (var (url, title) in deduped)
         {
             try
             {
+                // Fetch individual job page text via page-fetcher-api
+                var fetchResponse = await _pageFetcher.FetchAsync(new FetchPageRequest
+                {
+                    Url = url,
+                    WaitFor = "domcontentloaded",
+                    CallerService = "cv-search-job"
+                }, ct);
+
+                if (!fetchResponse.Success || string.IsNullOrWhiteSpace(fetchResponse.Text))
+                {
+                    _logger.LogWarning("Session {SessionId}: fetch failed for {Url} — {Error}", session.Id, url, fetchResponse.Error);
+                    continue;
+                }
+
+                var jobText = fetchResponse.Text;
+
+                // Keyword pre-filter: skip LLM call if no CV keyword appears in the job page text
+                if (cvKeywords.Count > 0 &&
+                    !cvKeywords.Any(k => jobText.Contains(k, StringComparison.OrdinalIgnoreCase)))
+                {
+                    _logger.LogInformation(
+                        "Session {SessionId}: pre-filter skip | {Url} | no CV keyword found in job text",
+                        session.Id, url);
+                    continue;
+                }
+
                 var matchRequest = new MatchJobRequest
                 {
                     CvDocumentId = session.CvDocumentId,
                     JobUrl = url,
+                    // Pre-fetched text passed directly so cv-matcher-api skips re-fetching the page
+                    JobDescription = jobText,
                     // User already gave GDPR consent when they clicked the one-time job search link
                     GdprConsent = true
                 };
@@ -182,7 +216,7 @@ public sealed class CvSearchJobTask : IJobTask
                     SessionId = session.Id,
                     ProviderName = GuessProvider(url, providers),
                     JobUrl = url,
-                    JobTitle = matchResult.Summary.Split('.').FirstOrDefault()?.Trim() ?? "Job",
+                    JobTitle = matchResult.Summary.Split('.').FirstOrDefault()?.Trim() ?? title,
                     JobText = string.Empty,
                     Score = matchResult.Score,
                     ResultJson = JsonSerializer.Serialize(matchResult, new JsonSerializerOptions(JsonSerializerDefaults.Web)),

Jobs/cv-search-job/cv-search-job.csproj

@@ -13,7 +13,6 @@
     <PackageReference Include="Microsoft.Extensions.Hosting" />
     <PackageReference Include="Microsoft.EntityFrameworkCore.SqlServer" />
     <PackageReference Include="Refit.HttpClientFactory" />
-    <PackageReference Include="Microsoft.Playwright" />
   </ItemGroup>
 
   <ItemGroup>
@@ -26,6 +25,7 @@
     <ProjectReference Include="..\..\Apis\cv-search-data\cv-search-data.csproj" />
     <ProjectReference Include="..\..\Apis\common\common.csproj" />
     <ProjectReference Include="..\..\Apis\email-data\email-data.csproj" />
+    <ProjectReference Include="..\..\Apis\page-fetcher-api-models\page-fetcher-api-models.csproj" />
     <ProjectReference Include="..\..\Helpers\startup-helpers\startup-helpers.csproj" />
     <ProjectReference Include="..\job-scheduler\job-scheduler.csproj" />
   </ItemGroup>

docker-compose/docker-compose.yml

@@ -72,6 +72,9 @@ services:
       - RagApi__BaseUrl=${RagApi__BaseUrl:-http://rag-api:8080}
       - RagApi__InternalApiKey=${RagApi__InternalApiKey:-}
 
+      - PageFetcherApi__BaseUrl=${PageFetcherApi__BaseUrl:-http://page-fetcher-api:8080}
+      - PageFetcherApi__InternalApiKey=${PageFetcherApi__InternalApiKey:-}
+
       - Ai__Provider=${Ai__Provider:-OpenAI}
       - Ai__OpenAI__ApiKey=${Ai__OpenAI__ApiKey:-}
       - Ai__OpenAI__ChatModel=${Ai__OpenAI__ChatModel:-gpt-4o-mini}
@@ -266,6 +269,9 @@ services:
       - EmailApi__BaseUrl=${EmailApi__BaseUrl:-http://email-api:8080}
       - EmailApi__InternalApiKey=${EmailApi__InternalApiKey:-}
 
+      - PageFetcherApi__BaseUrl=${PageFetcherApi__BaseUrl:-http://page-fetcher-api:8080}
+      - PageFetcherApi__InternalApiKey=${PageFetcherApi__InternalApiKey:-}
+
       - FileStorage__Path=${FileStorage__Path:-Files}
 
       - JobSearch__Enabled=${JobSearch__Enabled:-true}
@@ -293,6 +299,38 @@ services:
     labels:
       - "com.centurylinklabs.watchtower.enable=true"
 
+  page-fetcher-api:
+    image: registry.easysoft.ro/apps/myai-page-fetcher-api:${IMAGE_TAG:-staging}
+    container_name: myai-page-fetcher-api
+    environment:
+      - ASPNETCORE_ENVIRONMENT=${ASPNETCORE_ENVIRONMENT:-Staging}
+      - ASPNETCORE_URLS=${ASPNETCORE_URLS:-http://+:8080}
+      - APP_ENVIRONMENT_NAME=${APP_ENVIRONMENT_NAME:-myai.staging}
+
+      - Database__Host=${Database__Host:-sqlserver}
+      - Database__Port=${Database__Port:-1433}
+      - Database__Name=${Database__Name:-MyAiDb}
+      - Database__User=${Database__User:-sa}
+      - Database__Password=${Database__Password:-}
+      - Database__TrustServerCertificate=${Database__TrustServerCertificate:-true}
+
+      - InternalApi__ApiKey=${PageFetcherApi__InternalApiKey:-}
+      - InternalApi__RequireApiKey=true
+
+      - SerilogEmail__From=${SerilogEmail__From:-}
+      - SerilogEmail__To=${SerilogEmail__To:-}
+      - SerilogEmail__Host=${SerilogEmail__Host:-}
+      - SerilogEmail__Port=${SerilogEmail__Port:-587}
+      - SerilogEmail__UserName=${SerilogEmail__UserName:-}
+      - SerilogEmail__Password=${SerilogEmail__Password:-}
+    volumes:
+      - ${LOGS_PATH:-/opt/myai/logs}/page-fetcher-api:/app/logs
+    networks:
+      - myai-network
+    restart: unless-stopped
+    labels:
+      - "com.centurylinklabs.watchtower.enable=true"
+
   web:
     image: registry.easysoft.ro/apps/myai-web:${IMAGE_TAG:-staging}
     container_name: myai-web

myAi.sln

@@ -63,6 +63,12 @@ Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "email-api", "Apis\email-api
 EndProject
 Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "email-data", "Apis\email-data\email-data.csproj", "{C1D2E3F4-A5B6-4789-CDEF-012345678ABC}"
 EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "page-fetcher-api-models", "Apis\page-fetcher-api-models\page-fetcher-api-models.csproj", "{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}"
+EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "page-fetcher-data", "Apis\page-fetcher-data\page-fetcher-data.csproj", "{06F803CD-329D-40C2-B62D-0F14E137D3C7}"
+EndProject
+Project("{FAE04EC0-301F-11D3-BF4B-00C04F79EFBC}") = "page-fetcher-api", "Apis\page-fetcher-api\page-fetcher-api.csproj", "{FC5A722A-7B12-459E-AB9F-0A724797783E}"
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU
@@ -357,6 +363,42 @@ Global
 		{C1D2E3F4-A5B6-4789-CDEF-012345678ABC}.Release|x64.Build.0 = Release|Any CPU
 		{C1D2E3F4-A5B6-4789-CDEF-012345678ABC}.Release|x86.ActiveCfg = Release|Any CPU
 		{C1D2E3F4-A5B6-4789-CDEF-012345678ABC}.Release|x86.Build.0 = Release|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Debug|x64.ActiveCfg = Debug|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Debug|x64.Build.0 = Debug|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Debug|x86.ActiveCfg = Debug|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Debug|x86.Build.0 = Debug|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Release|Any CPU.Build.0 = Release|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Release|x64.ActiveCfg = Release|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Release|x64.Build.0 = Release|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Release|x86.ActiveCfg = Release|Any CPU
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B}.Release|x86.Build.0 = Release|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Debug|x64.ActiveCfg = Debug|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Debug|x64.Build.0 = Debug|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Debug|x86.ActiveCfg = Debug|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Debug|x86.Build.0 = Debug|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Release|Any CPU.Build.0 = Release|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Release|x64.ActiveCfg = Release|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Release|x64.Build.0 = Release|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Release|x86.ActiveCfg = Release|Any CPU
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7}.Release|x86.Build.0 = Release|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Debug|x64.ActiveCfg = Debug|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Debug|x64.Build.0 = Debug|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Debug|x86.ActiveCfg = Debug|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Debug|x86.Build.0 = Debug|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Release|Any CPU.Build.0 = Release|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Release|x64.ActiveCfg = Release|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Release|x64.Build.0 = Release|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Release|x86.ActiveCfg = Release|Any CPU
+		{FC5A722A-7B12-459E-AB9F-0A724797783E}.Release|x86.Build.0 = Release|Any CPU
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE
@@ -385,6 +427,9 @@ Global
 		{BE44B4EB-9AB9-4D81-A9BF-5CF2832BEEE5} = {A9B8C7D6-E5F4-4321-ABCD-FEDCBA987654}
 		{434119EA-2FFC-4433-9B8E-1E6D94006413} = {0FE6558F-2157-47F2-A835-558416CE0E2B}
 		{C1D2E3F4-A5B6-4789-CDEF-012345678ABC} = {D4E5F6A7-B8C9-4012-3456-789ABCDEF012}
+		{4F1A669E-C8AF-428F-87E7-3E0A213DD20B} = {A9B8C7D6-E5F4-4321-ABCD-FEDCBA987654}
+		{06F803CD-329D-40C2-B62D-0F14E137D3C7} = {D4E5F6A7-B8C9-4012-3456-789ABCDEF012}
+		{FC5A722A-7B12-459E-AB9F-0A724797783E} = {0FE6558F-2157-47F2-A835-558416CE0E2B}
 	EndGlobalSection
 	GlobalSection(ExtensibilityGlobals) = postSolution
 		SolutionGuid = {6246A67B-299E-4E64-8DBE-1A66771E7C67}