Getting started

Get started with jobContext

Three ways in, all backed by the same memory. Pick the one that fits how you work; you can add the others later, and everything stays in sync.

Fastest Cloud + your AI Sign in, add a connector. About a minute. Most private Desktop app Native app, local data, no account. Companion Mobile (iOS beta) Capture jobs from your phone's share sheet.

Cloud: connect your AI in about a minute

The hosted workspace at jobcontext.ai. No installs, no config files, no infrastructure.

Sign in with Microsoft

Sign in with any Microsoft account. Your private, isolated workspace is created on first login; nobody else's data touches yours.

Add the connector to your AI client

In Claude.ai, Cursor, or VS Code, add a remote MCP server pointing at:

https://jobcontext.ai/mcp

Your client opens a sign-in window and OAuth handles the rest: no keys to copy, no JSON to edit. The client discovers all 12 domain tools automatically.

Or skip the connector entirely: WebMCP

The dashboard is itself an agent surface. Open it in ChatGPT desktop's built-in browser while signed in and the in-page agent discovers every tool via document.modelContext, with no setup at all. Chrome (behind an origin trial) and Edge's preview gain the same tools as their WebMCP support lands.

Bootstrap your workspace

In your AI chat, say:

"Run workspace setup and walk me through it."

The workspace tool creates your whole data tree with zero manual setup. Then drop in your resume and start talking: log contacts, assess postings, generate documents. It remembers all of it next session.

Desktop: the whole platform as a native app

No terminal, no Python, no account. Local SQLite; everything stays on your machine.

Download and install

Grab the newest desktop-v* build from the releases page: a signed & notarized .dmg for macOS (Apple Silicon and Intel), an Authenticode-signed installer for Windows, and .AppImage/.deb for Linux. Updates install themselves.

Open it. That's the setup

The app runs the full server locally and opens the same dashboard the cloud serves. Chat with an embedded AI over your own data: bring your own OpenAI or Anthropic key, or point it at a local Ollama model and run with no keys and no cloud at all.

Optional: connect clients and sync

One-click MCP connect wires Claude Desktop, VS Code, or Cursor to your local server from the Settings screen. Link your cloud workspace and desktop ⇆ cloud sync keeps both sides current and feeds the mobile app.

Mobile: capture from your phone

The iOS companion, in TestFlight beta. Desktop creates knowledge; mobile captures reality.

Join the beta

The app is in TestFlight; see mobile/README.md for the current beta link and screens.

Paste an API key

Create a personal access token on the dashboard's API Keys tab, paste it once into the app's Settings, and it lives in your device keychain. No separate sign-in to go stale while the app sits unopened.

Share postings straight into your pipeline

See a role in Safari or LinkedIn? Share it to jobContext. Pages are extracted on your phone, so it reads postings that block datacenter IPs. Triage the queue and log wellbeing check-ins from anywhere; everything syncs back to desktop and cloud.

Prefer to run it yourself?

It's open source (MIT). Clone, create a venv, and run: stdio MCP for local clients or the full HTTP server with the dashboard.

git clone https://github.com/JustLikeFrank3/jobContextMCP
cd jobContextMCP
python3 -m venv .venv && .venv/bin/pip install -r requirements.txt
.venv/bin/python server.py          # stdio MCP server
# or the HTTP server: dashboard + REST + MCP Streamable HTTP at /mcp

Per-client walkthroughs (VS Code, Claude Desktop, ChatGPT desktop, Cursor, Windsurf), Docker mode, and deployment guides live in the repo: docs/client-setup.md and docs/local-development.md.

Ready to connect?

Open dashboard Download desktop