Engram
Log inGet started

Getting Started

Get Engram running and deploy your first guideline in under 10 minutes.

1. Create Your Workspace

SaaS: Sign up at app.getengram.io with email or GitHub/Google OAuth.

Self-hosted: Deploy with Docker — see the Self-Hosted Setup guide. Then open your dashboard and create an account.

Once signed in, create a workspace with a name (your company) and a slug (URL identifier, e.g. acme).

2. Connect Your Sources

Sources are the raw material Engram uses to understand your company. Navigate to Settings > Sources in the dashboard, or use the CLI:

# Install the CLI
bun install -g @engram/cli

# Connect sources
engram init
engram sources add github --org your-org
engram sources add website https://www.yourcompany.com
engram sources add notion

Engram scans your GitHub repos for existing rules files (CLAUDE.md, .cursor/rules, etc.), crawls your website for brand voice, and ingests Notion pages for documentation.

3. Run Gap Analysis

engram analyze

Or click Run Analysis in the dashboard. Engram analyzes all ingested content and produces:

  • Can create — topics with enough source material to generate a guideline
  • Issues found — contradictions, drift, or quality problems in existing rules
  • Missing critical — essential topics with no source material

4. Generate Guidelines

Review the suggested topics. Confirm the ones you want, dismiss the rest:

engram topics confirm "API Conventions"
engram topics confirm "Brand Voice"
engram generate "API Conventions"
engram generate "Brand Voice"

Each generates a guideline document from your source material. Domain owners review and approve them in the dashboard (status: needs-review → active).

5. Deploy

Once guidelines are active, deploy them to your repos:

engram deploy

This compiles guidelines into the right format for each AI tool (Claude Code, Cursor, Copilot, etc.) and creates PRs in your connected repos via the GitHub App.

For non-code use cases (sales, marketing, support), your team accesses context via the REST API — configure role-specific AI assistants with the Engram API baked in.

What's Next

Next
Deployment Models