gcontextgcontext.ai← Back

Build context like you build code

$ curl -LsSf gcontext.ai/gcontext/install.sh | sh
GitHub
Works withClaude CodeCursorCodexpi.dev

See it in action

Context modules give your agent structured files it can navigate by itself.

A real session, replayed. Read the full unedited transcript.

Get started

Three copy-pastes to a working agent

One command, two lines in .env, one prompt. Your agent does the rest.

  1. 1

    Install and initialize

    $ curl -LsSf https://gcontext.ai/gcontext/install.sh | sh$ gcontext init

    The only commands you have to run. From here on, your agent operates the workspace itself.

  2. 2

    Put a key in .env

    SUPABASE_URL=https://your-project.supabase.coSUPABASE_SECRET_KEY=sb_secret_...

    Modules only ever name the variables. The values stay in the gitignored .env.

  3. 3

    Say the prompt

    Create a supabase integration module for our Supabase project andload it into the workspace. The keys are in .env as SUPABASE_URLand SUPABASE_SECRET_KEY.

    Your agent (Claude Code, Cursor, Codex) writes the module files, declares the secrets by name, and loads it.

  4. 4

    Ask a real question in a fresh session

    How many users do we have?

    A session that has never seen your project follows the index, calls the API with the key from .env, and answers with the live row count.

The prompt in step 3 runs against a real model before every release; if it stops working, the release does not ship.
Read the full walkthrough · See a real recorded session

Your context is just files

No vector database. No prompt templates. A file tree your agent reads top-down.

context/
├─ llms.txt # the index the agent reads first
├─ integrations/
│ └─ postgres/
│ ├─ info.md
│ └─ module.yaml
├─ workflows/
│ └─ deploy/
│ ├─ info.md
│ └─ module.yaml
└─ tasks/
└─ sprint/
├─ info.md
└─ module.yaml

See it work

Hover or tap a panel to take a closer look.

No context vs. structured modules, side by side.

no contextno memory
Chat 1

Investigate support ticket SUP-123

I don't have access to that ticket. What system is it tracked in? What's the issue?

It's in Linear. User hasn't received payment, probably Supabase. Here are the logs: [14:31:02] ERR payment.webhook timeout [14:31:03] ERR edge_fn: payment-process

Based on those logs, the webhook timed out. Likely a Supabase edge function failure. Check payment-process errors around 14:31.

new chat
Chat 2

What's the status of SUP-123?

I don't have context for that ticket.

Used by

MAATmaatapp.comSoundmuraisoundmurai.comBleak AIbleakai.com

“It has transformed how I approach solving support tasks. From a boring task I was always procrastinating into something actually fun to do.”

Stefano Saitta·CTO, MAAT·

“Our gcontext Cloud agent helps us build posts for insights.soundmurai.com. It pulls global live-music data from our database to generate the insights.”

Eduard Aymerich·Co-Founder, Soundmurai·

Onboard a teammate in one clone

Joining an existing workspace takes zero setup and zero new concepts. Your new hire’s agent knows the codebase before they do.

$ git clone your-repo && cd your-repo
$ cp .env.example .env # if present, their own credentials
$ claude # the workspace tells the agent the rest

Their agent reads the same modules yours does: the schema notes, the deploy runbook, the gotchas your team already paid to learn. Nobody hand-maintains it: agents update modules as a side effect of doing work, and humans review the diffs in PRs.

Your workspace, your rules

Open source and local. Your context lives in your repo as files: portable, yours to keep.

CLI

Open source

Build and sync context modules from your terminal. Free forever.

$ curl -LsSf gcontext.ai/gcontext/install.sh | sh
View on GitHub →

gcontext Cloud

Coming soon

Everything in the CLI, plus a visual editor, AI chat with your context, and managed infrastructure.