Add startup.jobs extractor support (#279)

* Add startup.jobs extractor support

* Harden startup.jobs extractor inputs

* Wire startupjobs into Docker and CI

* Tighten startupjobs review follow-ups

* fix: publish ghcr during release workflow

* feat: add startupjobs max jobs configuration and update related tests

orchestrator/src/client/pages/OrchestratorPage.test.tsx

@@ -753,6 +753,7 @@ describe("OrchestratorPage", () => {
         gradcrackerMaxJobsPerTerm: 150,
         ukvisajobsMaxJobs: 150,
         adzunaMaxJobsPerTerm: 150,
+        startupjobsMaxJobsPerTerm: 150,
         jobspyCountryIndeed: "united kingdom",
         searchCities: "United Kingdom",
       });

orchestrator/src/client/pages/orchestrator/AutomaticRunTab.tsx

@@ -182,6 +182,7 @@ export const AutomaticRunTab: React.FC<AutomaticRunTabProps> = ({
 
     const rememberedRunBudget =
       settings?.jobspyResultsWanted?.value ??
+      settings?.startupjobsMaxJobsPerTerm?.value ??
       settings?.adzunaMaxJobsPerTerm?.value ??
       settings?.gradcrackerMaxJobsPerTerm?.value ??
       settings?.ukvisajobsMaxJobs?.value ??

orchestrator/src/client/pages/orchestrator/automatic-run.test.ts

@@ -52,6 +52,17 @@ describe("automatic-run utilities", () => {
     expect(cap).toBeLessThanOrEqual(750);
   });
 
+  it("assigns a dedicated startupjobs max-jobs limit", () => {
+    const limits = deriveExtractorLimits({
+      budget: 120,
+      searchTerms: ["backend", "platform"],
+      sources: ["startupjobs"],
+    });
+
+    expect(limits.startupjobsMaxJobsPerTerm).toBeGreaterThan(0);
+    expect(limits.startupjobsMaxJobsPerTerm).toBeLessThanOrEqual(120);
+  });
+
   it("returns zero estimate when no search terms are provided", () => {
     const estimate = calculateAutomaticEstimate({
       values: {
@@ -112,4 +123,21 @@ describe("automatic-run utilities", () => {
     expect(estimate.discovered.cap).toBeGreaterThan(0);
     expect(estimate.discovered.cap).toBeLessThanOrEqual(120);
   });
+
+  it("includes startupjobs in estimate caps using the shared term budget", () => {
+    const estimate = calculateAutomaticEstimate({
+      values: {
+        topN: 10,
+        minSuitabilityScore: 50,
+        searchTerms: ["backend", "platform"],
+        runBudget: 120,
+        country: "united kingdom",
+        cityLocations: [],
+      },
+      sources: ["startupjobs"],
+    });
+
+    expect(estimate.discovered.cap).toBeGreaterThan(0);
+    expect(estimate.discovered.cap).toBeLessThanOrEqual(120);
+  });
 });

orchestrator/src/client/pages/orchestrator/automatic-run.ts

@@ -66,6 +66,7 @@ export interface ExtractorLimits {
   gradcrackerMaxJobsPerTerm: number;
   ukvisajobsMaxJobs: number;
   adzunaMaxJobsPerTerm: number;
+  startupjobsMaxJobsPerTerm: number;
 }
 
 export function deriveExtractorLimits(args: {
@@ -82,6 +83,7 @@ export function deriveExtractorLimits(args: {
   const includesUkVisaJobs = args.sources.includes("ukvisajobs");
   const includesAdzuna = args.sources.includes("adzuna");
   const includesHiringCafe = args.sources.includes("hiringcafe");
+  const includesStartupJobs = args.sources.includes("startupjobs");
 
   const weightedContributors =
     (includesIndeed ? termCount : 0) +
@@ -90,7 +92,8 @@ export function deriveExtractorLimits(args: {
     (includesGradcracker ? termCount : 0) +
     (includesUkVisaJobs ? 1 : 0) +
     (includesAdzuna ? termCount : 0) +
-    (includesHiringCafe ? termCount : 0);
+    (includesHiringCafe ? termCount : 0) +
+    (includesStartupJobs ? termCount : 0);
 
   if (weightedContributors <= 0) {
     return {
@@ -98,6 +101,7 @@ export function deriveExtractorLimits(args: {
       gradcrackerMaxJobsPerTerm: budget,
       ukvisajobsMaxJobs: budget,
       adzunaMaxJobsPerTerm: budget,
+      startupjobsMaxJobsPerTerm: budget,
     };
   }
 
@@ -109,6 +113,7 @@ export function deriveExtractorLimits(args: {
     gradcrackerMaxJobsPerTerm: perUnit,
     ukvisajobsMaxJobs: Math.min(budget, perUnit + remainder),
     adzunaMaxJobsPerTerm: perUnit,
+    startupjobsMaxJobsPerTerm: perUnit,
   };
 }
 
@@ -173,6 +178,7 @@ export function calculateAutomaticEstimate(args: {
   const hasGlassdoor = sources.includes("glassdoor");
   const hasAdzuna = sources.includes("adzuna");
   const hasHiringCafe = sources.includes("hiringcafe");
+  const hasStartupJobs = sources.includes("startupjobs");
   const limits = deriveExtractorLimits({
     budget: values.runBudget,
     searchTerms: values.searchTerms,
@@ -191,9 +197,17 @@ export function calculateAutomaticEstimate(args: {
   const hiringCafeCap = hasHiringCafe
     ? limits.jobspyResultsWanted * termCount
     : 0;
+  const startupJobsCap = hasStartupJobs
+    ? limits.startupjobsMaxJobsPerTerm * termCount
+    : 0;
 
   const discoveredCap =
-    jobspyCap + gradcrackerCap + ukvisaCap + adzunaCap + hiringCafeCap;
+    jobspyCap +
+    gradcrackerCap +
+    ukvisaCap +
+    adzunaCap +
+    hiringCafeCap +
+    startupJobsCap;
   const discoveredMin = Math.round(discoveredCap * 0.35);
   const discoveredMax = Math.round(discoveredCap * 0.75);
   const processedMin = Math.min(values.topN, discoveredMin);

orchestrator/src/client/pages/orchestrator/usePipelineControls.ts

@@ -181,11 +181,13 @@ export function usePipelineControls(
       );
       const hasAdzuna = compatibleSources.includes("adzuna");
       const hasHiringCafe = compatibleSources.includes("hiringcafe");
+      const hasStartupJobs = compatibleSources.includes("startupjobs");
       const serializedCities = serializeCityLocationsSetting(
         values.cityLocations,
       );
       const searchCities =
-        (hasJobSpySite || hasAdzuna || hasHiringCafe) && serializedCities
+        (hasJobSpySite || hasAdzuna || hasHiringCafe || hasStartupJobs) &&
+        serializedCities
           ? serializedCities
           : formatCountryLabel(values.country);
       await api.updateSettings({
@@ -194,6 +196,7 @@ export function usePipelineControls(
         gradcrackerMaxJobsPerTerm: limits.gradcrackerMaxJobsPerTerm,
         ukvisajobsMaxJobs: limits.ukvisajobsMaxJobs,
         adzunaMaxJobsPerTerm: limits.adzunaMaxJobsPerTerm,
+        startupjobsMaxJobsPerTerm: limits.startupjobsMaxJobsPerTerm,
         jobspyCountryIndeed: values.country,
         searchCities,
       });

orchestrator/src/client/pages/orchestrator/utils.test.ts

@@ -17,6 +17,10 @@ describe("orchestrator utils", () => {
     expect(getEnabledSources(withoutKey)).not.toContain("adzuna");
   });
 
+  it("enables startupjobs without credentials", () => {
+    expect(getEnabledSources(createAppSettings())).toContain("startupjobs");
+  });
+
   it("counts processing jobs in ready and discovered tabs", () => {
     const jobs = [
       createJob({ id: "ready", status: "ready", closedAt: null }),

orchestrator/src/client/pages/orchestrator/utils.ts

@@ -195,6 +195,10 @@ export const getEnabledSources = (
       enabled.push(source);
       continue;
     }
+    if (source === "startupjobs") {
+      enabled.push(source);
+      continue;
+    }
     if (
       source === "indeed" ||
       source === "linkedin" ||

orchestrator/src/server/config/demo-defaults.data.ts

@@ -254,6 +254,7 @@ export const DEMO_SOURCE_BASE_URLS: Record<JobSource, string> = {
   ukvisajobs: "https://www.ukvisajobs.com",
   adzuna: "https://www.adzuna.com",
   hiringcafe: "https://hiring.cafe",
+  startupjobs: "https://startup.jobs",
   manual: "https://example.com",
 };