Find your shortlist
Ask for a role and location. Your agent checks available sources and presents jobs for you to choose.
JobSpidey for AI agents
Your AI agent. Your next application.
Bring job search, tailored resumes and cover letters, usable PDFs, and application tracking into your agent’s workflow with Model Context Protocol (MCP).
JobSpidey prepares and tracks. Your agent handles the employer website and submission, with your direction.
One connected workflow
From shortlist to a ready application pack
Prepare selected jobs
Your agent prepares a tailored resume and cover letter, follows progress, and retrieves the PDFs.
Apply and keep track
Your agent handles the employer website. It can report the outcome back to your JobSpidey tracker.
“Find data engineering jobs in Auckland. Show me a shortlist, prepare documents for the two jobs I choose, and give me the PDFs.”
Example request; it does not start a search or spend credits from this page.
Connect without copying a secret
Set up your agent
Your client starts the connection. Sign in or create an account on JobSpidey, review the requested permissions and daily generation budget, then return to your agent.
Setup commands will appear here when this environment has an enabled endpoint. No account password, API key or access token belongs in your conversation.
Codex desktop and CLI
- Add the configured remote server on the Codex host you use.
- Run OAuth login and complete the JobSpidey consent screen. For a pre-registered pilot, use the public client ID supplied by the pilot operator with
--oauth-client-id; this ID is not a secret. - Check
codex mcp listor/mcp, then ask your agent to readget_account_contextbefore searching.
Codex desktop and CLI share MCP configuration on the same host. See the official Codex MCP guide for current settings and OAuth client registration.
Search and tracking can work before you supply a resume. Document preparation requires a usable saved resume or a permitted one-off text override, and uses normal account credits within the connection budget.
You control the connection
Consent separates candidate facts, resume-text access, search, document generation, document reads and optional tracker writes. Generated documents contain candidate information even when the agent cannot read your master resume.
Replay or reuse does not charge again. Refreshing expired PDF links does not regenerate documents. A partial package is labeled partial, so your agent can see what is missing.
Manage Connected AI agents to review permissions, budget usage and last use, or disconnect. Disconnect blocks new access; already admitted generation may finish in your account.
Coverage follows available sources
Search availability depends on country, source support and current service limits. English setup guidance is prepared first. Worldwide discovery does not promise jobs from every country.
Ask your connected agent for the current countries and search sources from get_account_context before starting a search.
Version 1 contract
Tools for the whole preparation journey
Availability follows this environment’s rollout settings and your connection permissions. Searches use stable handles; mutations use a request ID so retrying an accepted request reuses the same work.
| Tool | What your agent can do | Permission |
|---|---|---|
get_account_context | Check permissions, credits, resume readiness, supported search sources and templates. | Connection access |
get_candidate_context | Read permitted candidate facts and their sources. Resume text needs separate permission. | profile:read; resume:read for resume text |
search_jobs | Start a search with a request ID and keep its durable search handle. | jobs:search + jobs:read |
list_search_results | Page through brief, ranked results from that exact search. | jobs:read |
get_job | Read a selected job’s description, source and posting link. | jobs:read |
load_more_jobs | Extend the same search where its source supports more results. | jobs:search + jobs:read |
prepare_application | Prepare a selected or custom job using your normal account entitlements. | documents:generate |
get_operation | Follow search or preparation progress and the server’s retry guidance. | Permission for the operation |
get_application_package | Retrieve document readiness, readable text and expiring PDF download links. | applications:read + documents:read |
list_applications | Read your Track Jobs summaries and package readiness. | applications:read |
record_application_outcome | Record an agent-reported outcome after your agent’s external submission. | applications:write (optional) |
Keep your next step clear
Review the job and documents before submission. Outcomes from your agent are recorded as agent-reported; JobSpidey does not contact the employer or verify submission.