Best for
- Job seekers who want to self-host and control their career data.
- Users who want to record jobs or interview questions without leaving an AI chat.
- Users of AI clients such as Claude Desktop, OpenClaw, or Hermes.
J
JobSync MCP Server
Community
A self-hosted AI assistant for job search and application tracking
- Category
- Other #96 of 230
- Stars
- ★ 1.4k Very popular
- Transport
- Streamable HTTP
- Runtime
- Node.js · Docker
- Credentials
- API key / credential required
- License
- MIT
- Last commit
57FMRS · C
The source presents JobSync MCP as a practical bridge between AI clients and a self-hosted job-tracking application, with local Ollama support as a privacy-oriented option. No concrete MCP tool identifiers or complete permission list are provided, so the tool list is left empty. Protect the access token and review AI-generated results before saving.
Strongest · Documentation 14/20 Weakest · Reliability 8/20
Why each score
- Reliability 8/20
- The README describes a built-in MCP endpoint, token authentication, and several write-oriented use cases, but provides no server manifest, tool inventory, source code, CI, or tests. MCP initialization, tool-list handshake, actual parameters, error handling, and consistency between declarations and behavior are therefore unverifiable. The normal path is plausible, but static uncertainty requires a substantial deduction.
- Security and permissions 13/20
- The evidence indicates Bearer-token authentication, write operations for jobs and question-bank entries, and claimed user approval before saving. The in-app assistant is described as having narrow data access, and pasted job postings are treated as untrusted text. Points are deducted because the missing source and manifest prevent verification of token storage, permission scope, expiry and revocation, transport protection, cloud-AI data flows, session isolation, CSRF defenses, and whether confirmation is enforced. No demonstrated red-line issue is present.
- Maintenance 10/20
- The project is MIT-licensed, not archived, and provides an update path plus links to contribution and conduct guidelines, indicating a basic maintenance surface. Points are deducted because the supplied material contains no commit history, release cadence, dependency-update evidence, security-response channel, maintainer accountability, or issue-response evidence. Stars and open issues do not add maintenance points.
- Documentation 14/20
- The README supplies Docker startup instructions, port and account guidance, MCP token generation, Claude Desktop and streamable-http configuration examples, provider information, tool-calling requirements, HTTP-network caveats, and some feature boundaries. Points are deducted for the absence of a manifest or per-tool parameter and return-value documentation, reviewable error examples, a compatibility matrix, concrete cost and data-retention details, and MCP troubleshooting or verification evidence.
- Setup experience 12/20
- Docker startup is clear, and the README includes token-generation steps and configuration examples for mainstream clients. It also explains the mcp-remote bridge, HTTPS/LAN HTTP behavior, and client restart requirement. Points are deducted because users must configure Docker, an account, a token, AI settings, and potentially a client bridge; there is no manifest, installation verification, platform compatibility matrix, or source-level execution evidence. The setup is usable but still manual and only partially verifiable.
Static review · not runListed 2026-08-14
Read the FMRS scoring method →Fit and risk
What it can accessUses the networkChanges third-party account data
Not for
- Users expecting the MCP server to provide full job search, task, or activity-management capabilities; the README does not document those MCP capabilities.
- Users unwilling to grant an AI client permission to write job applications or Question Bank entries.
- Users seeking a hosted third-party integration without deploying JobSync themselves.
Required permissions
- Access to the self-hosted JobSync /api/mcp endpoint.
- A Bearer access token generated under Settings > MCP Access.
- With user approval, the ability to add job applications and Question Bank entries and resolve or create related companies, titles, locations, and tags.
- When a default resume and sufficiently substantial job description are available, the ability to save a job-match result based on that resume.
Risks and side effects
- The access token is shown only once; exposure could allow others to access the MCP endpoint.
- AI-generated matching scores, recommendations, and written content require human review.
- Job-posting text is treated as untrusted input, but users should still inspect confirmation-card details before saving.
- Cloud AI providers introduce different privacy considerations from running models locally with Ollama.
Setup
Before you start
Runtime：Node.js · Docker
AUTH_SECRET
Self-hosted auth secret, auto-generated if unset; set manually with openssl rand -base64 32.
Other optional settings (2)
TZ
Self-hosted timezone variable (e.g. America/Edmonton); set on remote servers in docker-compose.yml to avoid activity time shifts.
NEXTAUTH_URL
Set in .env after updating a self-hosted instance; on homelab use the server IP instead of localhost.
Sign in to JobSync, open Settings > MCP Access, generate an access token, and copy the configuration snippet. Claude Desktop requires mcp-remote; add the configuration to mcpServers in claude_desktop_config.json and fully restart the client. Clients with native streamable-http support can connect directly to /api/mcp using an Authorization Bearer token. For a plain HTTP LAN address, mcp-remote may require the --allow-http argument.
claude_desktop_config.json
{"mcpServers":{"jobsync":{"command":"npx","args":["mcp-remote","http://<your-jobsync-url>/api/mcp","--header","Authorization: Bearer <your-token>"]}}}
Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.
.vscode/mcp.json
{
"servers": {
"jobsync": {
"command": "npx",
"args": [
"mcp-remote",
"http://<your-jobsync-url>/api/mcp",
"--header",
"Authorization: Bearer <your-token>"
]
}
}
}
Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).
Terminal
claude mcp add jobsync -- npx mcp-remote 'http://<your-jobsync-url>/api/mcp' --header 'Authorization: Bearer <your-token>'
Run it in a terminal; replace any <…> placeholders with your own values first.
Check that it works
Generate a token under Settings > MCP Access in JobSync, add the jobsync server to your MCP client (Claude Desktop bridges via npx mcp-remote), fully restart the client, then ask the agent to add a test job and confirm it appears in the app.
Troubleshooting
- Verify that JobSync is running and that the /api/mcp URL is reachable from the client.
- Verify that the Bearer token is complete, unexpired, and placed in the Authorization header.
- Claude Desktop must use mcp-remote and be fully quit and restarted after configuration changes.
- Add --allow-http when required for a non-HTTPS LAN URL.
- Confirm that JobSync has a selected AI model supporting structured output and tool calling.
Things to try
Once connected, you can ask your AI assistant things like:
- Add this job to JobSync: Senior Backend Engineer at Acme Corp, with this location and full job description...
- I just got asked this in an interview — add it to my Question Bank: 'How would you design a rate limiter?'
- Analyze this job description against my default resume and save the job match score
Use cases
Add job applications from chat while resolving or creating companies, titles, locations, and tags.
Record interview questions and answers in the Question Bank from chat.
Save a job-match score, recommendation, and analysis when a substantial job description and default resume are available.
Supported clients
Claude DesktopPartial support
OpenClaw
Hermes
Listed from the project's documentation, not tested by this site.
Overview
JobSync is a free, open-source, self-hosted job search assistant for managing applications, resumes, tasks, and job-search analytics. Its built-in MCP server lets compatible AI clients add job applications and Question Bank entries with user approval, and save a job-match analysis when the required resume and job-description conditions are met. AI features can run locally through Ollama or through several cloud model providers.
Similar servers
Hevy MCP
77 · B
Manage your Hevy workout data from AI assistants via MCP.
Firecrawl MCP Server
75 · B
Firecrawl's official MCP server for web search, scraping, and structured extraction for AI agents
BioMCP
75 · B
One binary. One grammar. Evidence from the biomedical sources you already trust.
Compartment
75 · B
Encrypted, fully offline long-term memory for AI agents
Source revision 2bd2284879d2 Data synced 2026-10-19 Read the FMRS scoring method