Lamr runs a Model Context Protocol (MCP) server. Connect it to Claude, ChatGPT, or Codex and ask it to screen a posting, check your pipeline, or draft tailored materials — right from the chat you're already in. Every credit rule and safety boundary from the app still applies.
The Model Context Protocol is an open standard for connecting AI assistants to external tools and data — instead of Lamr building a separate, one-off integration for Claude, another for ChatGPT, and another for Codex, it exposes one MCP server that any compatible client can discover and call the same way. Connecting it doesn't change what Lamr can do — it changes where you can ask for it.
No special syntax, no slash commands to memorize. A normal sentence is enough — the assistant resolves which tool to call and asks before anything destructive or credit-spending happens.
A few more that work as-is — click to copy:
Every tool maps to a real Lamr endpoint and carries an honest safety label. Reads never write, writes never delete, and anything destructive or credit-spending needs your assistant's own approval step before it runs.
list_applicationsFilter or search the pipeline by status, company, or role and get back stable IDs for every other call.
get_applicationRead one application's JD, notes, status, fit score, prep, and whether documents are ready.
create_applicationTrack a new opportunity from a pasted JD or URL. Never claims to have submitted it.
update_applicationChange status, notes, follow-up date, recruiter info, or which documents a job wants.
delete_applicationPermanently remove exactly one confirmed application. Always asks first.
get_candidate_profileRead target roles, location, and constraints — for staying consistent outside Lamr too.
get_master_resumeRead the structured tailoring source only when a task actually needs it.
get_account_usageRead tier, remaining credits, and billing-cycle state — never payment or provider identifiers.
screen_applicationRun and save Lamr's honest fit verdict, with reasons, before you invest further.
draft_application_materialsGenerate an editable, structured resume and cover-letter draft for human review.
compile_application_materialsCompile a reviewed draft into the typeset PDF and persist it to the application record.
Same server URL for every client. Requires an existing, invite-activated Lamr account — MCP authenticates as you, it doesn't get around the invite gate.
In Claude, go to Settings → Connectors (claude.ai, Desktop) and choose to add a custom connector by URL.
Use the URL below. Claude discovers Lamr's tools automatically over the Streamable HTTP transport.
Complete the OAuth sign-in with your Lamr account, then approve the connection. Claude can now use all 11 tools.
In ChatGPT, open Settings → Connectors and enable developer mode if it isn't already on.
Create a new connector using the URL below as the remote MCP endpoint.
Sign in with your Lamr account when prompted. ChatGPT can now call Lamr's tools mid-conversation.
Run codex mcp add lamr --url <URL> in your terminal, using the URL below.
Run codex mcp login lamr and complete the OAuth sign-in with your Lamr account.
Codex now has the same 11 tools — useful for scripting a pipeline check into your own workflow.
https://job-search-api-cqrv.onrender.com/mcp
A remote Model Context Protocol server that lets an AI assistant — Claude, ChatGPT, or Codex — read and act on your Lamr job-search workspace directly from a chat. It exposes the same application tracking, screening, and resume-tailoring capabilities as the web app through 11 structured tools, delegating every operation to Lamr's existing handlers so credit rules, validation, and tenant isolation stay on one code path.
An open standard for connecting AI assistants to external tools and data sources over a common protocol, instead of every app building a one-off integration for every assistant. A server like Lamr's exposes a fixed set of tools; any MCP-compatible client can discover and call them the same way.
Verified with Claude (claude.ai, Claude Desktop, and Claude Code), ChatGPT's developer-mode connectors, and Codex. Any client that supports the Streamable HTTP MCP transport and OAuth 2.1 can connect using the same server URL.
No. Lamr MCP tracks, screens, and drafts materials, but it never submits an application to an employer and never messages a recruiter on your behalf. Compiling a document is the furthest any tool goes — the apply step is always yours.
Connecting uses OAuth 2.1 with PKCE through Lamr's existing authentication provider — no password or API key is ever pasted into the assistant. Every tool call is authenticated as you and filtered to your own data, the same authorization Lamr's web app already enforces. Destructive actions like deleting an application still require the assistant to ask for confirmation before acting.
No separate pricing. Tool calls that generate AI content — screening a posting or drafting a resume — spend credits from the same monthly allowance as the web app, at the same cost per action. Read actions like listing applications or checking account usage are free.
Yes. Lamr MCP is not a way around the invite-gated signup — it authenticates as an existing, activated Lamr account. If you don't have one yet, join the waitlist below for an invite code first.
MCP is another door into the same workspace — everything below still runs through the same handlers, credits, and truth rules.
Private alpha
Leave your email and you'll get an invite the moment the doors open — one email, no spam.