Developer Docs & RFCs
This is the contract desk for the Job Pool ecosystem: JobDataPool API usage, OpenAPI and schema references, canonical dataset paths, publisher integration signals, responsible crawling policy, and the RFCs that explain why the system is split across several domains.
Surface Map
The topology is both a product boundary and an SEO boundary. Each site answers a different search intent, then hands users to the next surface when they need more depth.
mewannajob.comConsumer product for job seekers: ChatGPT apply plans, mobile quick search, desktop dashboards, Q&A, and outcome-first browsing.
jobdatapool.comCanonical infrastructure for developers: job data API docs, OpenAPI, JSON schemas, versioned datasets, publisher docs, and RFCs.
jobpool.liveTransparency layer for power users: reviewed downloads, scraper docs, staged submissions, review gates, leaderboards, and freshness signals.
datapool.workOperations layer for contributors: source submission, scraper orchestration, moderation, review tooling, and ingestion workflows.
For API Users
JobDataPool is the structured first stop for builders, researchers, aggregators, and downstream job products. Reach for the cache first, keep attribution intact, and make origin crawling the exception instead of the default.
/v1/jobs for current listings, with limit at or below 50./v1/sources before importing so your system knows the dataset repository, DVC pointer, and CSV origin./datasets/latest.csv when you need a bulk bootstrap, reproducible demo, analytics run, or offline QA fixture.url, apply_link, source_business_url, and ingest timestamps are part of the trust layer.curl -s "https://api.jobdatapool.com/v1/jobs?limit=25&country_code=US"
curl -s "https://api.jobdatapool.com/v1/sources"
curl -L "https://jobdatapool.com/datasets/latest.csv"| Endpoint | Use it for | Notes |
|---|---|---|
/v1/jobs | Current job listing queries | No auth in v1, 5 requests per 60s, 50 listings per request, 200 listings per 60s per IP/origin, limit capped at 50. |
/v1/sources | Canonical source catalog | Repository, DVC pointer, R2 CSV URL, schema, and endpoint references. |
/datasets/latest.csv | Reviewed dataset bootstrap | Redirects to the latest reviewed CSV snapshot for bulk jobs data workflows. |
/openapi.json | Contract automation | Use for SDKs, tests, integration checks, docs generation, and agent context. |
For Data Sources
If your site is the origin for job listings — a board, an ATS feed, a careers page, or a periodic dataset — you can plug into the pool through three documented integration paths. Reviewed scrapers, feed crawls by JobPoolBot, and direct dataset uploads all flow through the same staged review queue.
JobPoolBot to crawl on a paced schedule.job-listing.schema.json for backfills or periodic exports.JobPoolBot is the canonical data-source agent. It identifies as JobPoolBot/1.0 (+https://jobpool.live/jobpoolbot), honors robots.txt with Crawl-delay, prefers structured feeds, and routes ingested rows through the staged review queue before they appear in /v1/jobs.
For Job Board Publishers
Publisher integration should be transparent, RFC-compatible, and non-exclusive. The useful message is simple: your job board can point responsible consumers toward a structured JobDataPool feed before they repeatedly crawl origin pages.
robots.txt, sitemap, and search crawler behavior intact.Link header with rel="alternate" and type="application/json"./.well-known/, or an HTML metadata tag.# /robots.txt
User-agent: JobPoolBot
Allow: /
# Preferred structured cache:
# https://jobdatapool.com/publisher/exampleboard
# Try the cache/API before re-crawling origin pages.
Sitemap: https://example.com/sitemap.xmlMachine-readable Signals
These signals are small on purpose. A publisher should be able to add them without a platform migration, and a responsible data consumer should be able to find the structured feed without guessing.
Use the standard Link header first. Treat custom headers as optional hints for clients that know JobDataPool.
Link: <https://jobdatapool.com/publisher/exampleboard>; rel="alternate"; type="application/json"
X-JobDataPool-Canonical: https://jobdatapool.com/publisher/exampleboardAdd page-head hints for CMS templates, job detail pages, and crawlers already parsing HTML.
<link rel="alternate"
type="application/json"
href="https://jobdatapool.com/publisher/exampleboard">
<meta name="jobdatapool-canonical"
content="https://jobdatapool.com/publisher/exampleboard">Use one token across DNS, metadata, or /.well-known/ so publisher ownership can be confirmed without redesigning job pages.
jobdatapool-verification=abc123
/.well-known/jobdatapool-verification
<meta name="jobdatapool-verification" content="abc123">Expose a compact descriptor for publisher identity, freshness, usage context, and origin attribution.
{
"publisher": "ExampleBoard",
"origin": "https://example.com",
"canonical_feed": "https://jobdatapool.com/publisher/exampleboard",
"refresh_interval": "PT5M",
"allowed_usage": ["aggregation", "analytics", "search"]
}Access Policy
The positioning should stay standards-oriented and publisher-friendly. JobDataPool is a structured alternative and traffic offload layer, not a gatekeeper, authorization system, or exclusive access claim over public job data.
Developer FAQ
These are the questions this docs hub should answer before anyone has to spelunk through the repo or reverse-engineer the domain topology.
JobDataPool is the canonical infrastructure surface for Job Pool APIs, schemas, OpenAPI contracts, versioned datasets, documentation, and architecture RFCs.
Job board publishers can expose preferred access signals through robots comments, HTTP Link headers, HTML metadata, verification tokens, and feed descriptors.
No. JobDataPool is a cooperative structured alternative and traffic offload layer, not mandatory usage, authorization, or exclusive access to public job listings.
Agent Context
Developer agents should not reconstruct the ecosystem from stray page copy. The repo includes tool-specific entry files that point to one markdown instruction set for domain roles, SEO boundaries, API contracts, publisher integration, and implementation rules.
Markdown playbooks in docs/agent-instructions/ cover monthly DVC/R2 rollouts, the
/datasets/ catalog, RFC requirements, and a repo map. Start at
README.md in that folder (also linked from tool entry files).
Step-by-step checklist: jobpool-listings-r2 DVC pointers, public R2 artifacts, redirect-index
prefixes, jobdatapool.com env, and sibling apps (mewannajob, livejobpool).
AGENTS.md, CLAUDE.md, and .cursor/rules/jobdatapool.mdc all point agents at the same instruction index.
Architecture RFCs
The RFCs are still the durable record. They anchor the Job Pool data model, web topology, domain responsibilities, and implementation requirements while this docs hub gives developers the quick path in.
The foundational proposal for treating job listings as shared, structured, continuously refreshed data infrastructure.
The topology record for domain inventory, user journeys, cross-domain linking, and implementation requirements across the ecosystem.