Features / Project Context

Shared context for
every AI agent.

Your architecture, conventions, and key decisions — available to every AI coding session in the repo, automatically.

Write it once. Every AI tool that touches your codebase starts with the knowledge it needs.

Get started

project context

Your context

# Project Context

Next.js 14 with App Router

PostgreSQL via Prisma ORM

Trunk-based dev + feature flags

Deploy to Vercel (preview + prod)

Every AI session reads it

Claude Code, Cursor, Codex, Gemini CLI

Loaded via MCP automatically

How it works

Step 1

Create context

terminal

$ sfs project init

Detected: github.com/acme/web-app

Project context created.

Edit with: sfs project edit

Step 2

Write your knowledge

editor

$ sfs project edit

# Project Context

## Overview

## Architecture

## Conventions

## Key Decisions

Step 3

Every agent reads it

claude code

> What's the architecture?

[Reading project context via MCP...]

"This project uses Next.js 14 with App Router, PostgreSQL via Prisma ORM, and deploys to Vercel..."

How many times have you explained your architecture to an AI tool?

Every developer re-explains the same things: "We use PostgreSQL, not MySQL." "The API is FastAPI." "Don't touch migrations directly."

Project Context: explain once. Every AI session starts with that knowledge — no preamble, no repetition.

✓ Architecture decisions persisted across sessions

✓ Coding conventions enforced from day one

✓ API contracts and boundaries documented once

✓ Key decisions never lost to chat history

every session

Without Project Context

"We use PostgreSQL, not MySQL."

"The API is FastAPI, not Flask."

"Don't modify migrations directly."

Repeated every session, by every developer.

With Project Context

Already loaded. Zero prompting needed.

Team multiplier

New member onboarding

New developer joins, installs SessionFS, and every AI tool they use immediately knows your codebase conventions. No ramp-up period.

One source of truth

One document per repo, synced to the cloud. Update conventions once and every team member's AI sessions pick up the changes.

Consistent AI output

When every AI agent knows your conventions, generated code follows your patterns. Fewer review comments, less rework.

Manage from the dashboard

The Projects page in the dashboard shows all your repos with session counts, last activity, and quick access to edit project context.

CLI and dashboard stay in sync. Edit context from either — the latest version is always what your AI agents see.

dashboard / projects

acme/web-app

14 sessions

acme/api-gateway

8 sessions

acme/mobile-app

3 sessions

Context configured for all projects

Try it now

Two commands to give every AI agent your project knowledge.