Skip to content

Bring your own agent

Jobably works with the AI agent you already use. Connect it once and it can read your daily short list, search for jobs, tailor your resume, show you the draft, record what you applied to, and build you reports — all inside your account, all on the record.

There are two ways to connect, depending on the agent:

  • Chat apps (ChatGPT, claude.ai, Claude Desktop) — add Jobably as a connector and sign in. You approve the connection in Jobably. Nothing to copy.
  • Coding and command-line agents (Claude Code, Codex, GitHub Copilot, Cursor) — paste one command with a personal token.

The address is the same for every agent: https://mcp.jobably.com/mcp

  1. In ChatGPT, open Settings → Apps & Connectors → Advanced and turn on developer mode (your workspace admin may need to allow it).
  2. Back in Apps & Connectors, choose Create and enter the address https://mcp.jobably.com/mcp. Name it Jobably and choose OAuth.
  3. ChatGPT sends you to Jobably. Sign in if asked, check what it may do, and choose Allow.
  1. Open Settings → Connectors → Add custom connector.
  2. Enter the address https://mcp.jobably.com/mcp and choose Add.
  3. Choose Connect. Claude sends you to Jobably; sign in if asked and choose Allow.

Give it the address. If it supports OAuth for MCP servers it finds the sign-in on its own and sends you to Jobably to approve it.

The approval screen shows the agent’s name, where it will send you back to, and what it may do:

  • Read your job search — matches and scores, your short list, searches, the board, resumes and drafts, reports, and the activity record.
  • Act on it for you — move jobs on the board, record applications, run searches, tailor resumes (this uses your plan’s credits), and build lists and reports.

Untick the second to connect an agent that can only look. You can disconnect any agent at any time in Settings → Integrations → Connected agents.

Coding and command-line agents: one command

Section titled “Coding and command-line agents: one command”
  1. In the app, open Settings → Integrations → Personal access tokens and create a token. Give it read, and write if the agent should act.
  2. The Connect your AI assistant card right below shows the full command for your agent and operating system with the new token already filled in. Copy, paste, done.

The same commands, for reference:

Terminal window
claude mcp add jobably --transport http https://mcp.jobably.com/mcp \
--header "Authorization: Bearer YOUR_TOKEN"

Claude Code also supports sign-in: leave out the --header line and run /mcp inside Claude Code to connect through the approval screen instead.

Terminal window
codex mcp add jobably -- npx -y mcp-remote https://mcp.jobably.com/mcp \
--header "Authorization: Bearer YOUR_TOKEN"
Terminal window
code --add-mcp '{"name":"jobably","type":"http","url":"https://mcp.jobably.com/mcp","headers":{"Authorization":"Bearer YOUR_TOKEN"}}'
{
"mcpServers": {
"jobably": {
"type": "http",
"url": "https://mcp.jobably.com/mcp",
"headers": {
"Authorization": "Bearer YOUR_TOKEN"
}
}
}
}

You don’t need to set anything up in the app first. Connect your agent — signing in creates your Jobably account if you don’t have one — then ask:

“Set me up on Jobably. Here’s my resume.”

Attach or paste your resume. Your agent uploads it, and Jobably proposes a profile and a first search from it: your current titles, the level they sit at and where your current role is based, each with where it was read. Your agent shows you the proposal and asks what a resume can’t say — the lowest salary worth seeing, remote, hybrid or on-site, and any industries to favour or avoid. Say yes and it saves everything in one step. The first scan runs immediately; your first matches arrive within minutes, scored against your profile. Ask “How’s my Jobably setup going?” at any point and it tells you the next step.

Nothing is saved until you approve it, and everything stays editable in the app afterwards. The same quick start is in the app’s onboarding.

“What’s on my Jobably short list today? Tailor a resume for the best one and show me the draft.”

Then, once you have applied:

“I applied to the Acme role this morning — mark it applied.”

See Your agent’s workspace for the daily short list, lists, reports and the activity record.

Every tool says up front whether it is free or metered. Anything that spends your plan’s credits or changes your account needs the agent to pass confirm: true — a call without it is refused with an explanation, so a well-behaved agent asks you first. Chat apps also see which tools only read, so they can ask before the others.

Your agent never applies to a job for you. You apply; your agent records it.

Arguments are a plain JSON object. Two calls, shaped exactly right, cover the mistakes that actually happen:

{
"name": "add_accomplishments",
"arguments": {
"bank": [
"Negotiated $4.8M savings, 20% on $24M of capital equipment spend.",
"Cut sourcing cycle time 38% across a $40M program."
]
}
}

bank is a real JSON array — not a quoted string containing one — with one accomplishment per element. Commas stay INSIDE an entry: the first element above is one accomplishment, not three. Never pre-split text on commas; split on line breaks.

{
"name": "create_search",
"arguments": {
"name": "VP Engineering — remote, US",
"query": "\"VP Engineering\" OR \"VP of Engineering\"",
"location": "United States",
"experienceLevel": "5-10,10+",
"excludeTitles": ["recruiter"],
"remoteOnly": true,
"cadenceHours": 24,
"confirm": true
}
}

confirm is the boolean true, not the string "true". Numbers may arrive as strings and coerce ("24" means 24), but unparseable values error loudly rather than being replaced with a default. Ids (jobId, searchId, listId, reportId, operationId, artifactId, resumeId) always come from a prior call — never invent one.

Every tool in the reference carries an example call validated against the live schema in CI — copy its shape.

Signing in, creating tokens, approving other agents, model keys, billing and deleting the account happen only in the app, signed in as you. An agent can never create credentials or spend money outside the metered tools.

search_jobs runs one-off searches that are never scored against your profile and never saved. With forSomeoneElse: true, level and size filters are relaxed entirely — useful when a friend asks “can you look for staff-level roles for me?” without your rubric contaminating their results.

Built v1.1.0 · cec2567 · 2026-11-03 18:41 UTC