# Onramp

> AI-powered developer onboarding. Onramp indexes your codebase into a live architecture map and guides developers through their first PR — turning weeks of tribal knowledge gathering into days of self-serve productivity.

## Key Pages
- [Product Overview](https://onramp.app/) — What we do and how it works
- [Why Onramp](https://onramp.app/why-onramp) — Comparison vs coding agents
- [Pricing](https://onramp.app/#pricing) — Transparent, flat pricing
- [Documentation](https://onramp.app/docs) — Guides, API reference, self-hosting
- [Blog](https://onramp.app/blog) — Engineering insights and product updates
- [Security](https://onramp.app/security) — current security controls, audit status, and policies

## What We Do

Onramp solves the developer onboarding problem by turning registered GitHub repositories into interactive knowledge maps. Instead of new hires spending 3-6 months piecing together tribal knowledge from senior engineers, Onramp parses the actual source code, builds a live dependency graph, and creates guided onboarding paths with an AI mentor that answers questions with file and line references.

**Core capabilities:**
- **Live Architecture Map**: Visual, searchable dependency graph updated on every push
- **AI Codebase Mentor**: Ask questions in plain English, get answers grounded in your actual code with citations
- **Guided Learning Paths**: Structured onboarding plans tailored to each hire's background
- **First Issue Finder**: Curated, scope-appropriate first PRs for new developers
- **Architecture Drift Detection**: Flags when the actual architecture diverges from documented design
- **Time-to-First-PR Analytics**: Measure and improve onboarding velocity across teams

## Target Audience
- **Primary**: Engineering managers, CTOs, and VPs of Engineering at companies with 10-500+ developers
- **Secondary**: Developer experience teams, platform teams, onboarding program owners
- **Tertiary**: Individual developers joining new teams, open source contributors

## Key Differentiators
- **Parse once, answer forever**: Unlike coding agents that re-read the entire codebase on every query (burning tokens), Onramp indexes once and incrementally updates the knowledge graph on each push
- **Flat pricing, not per-seat**: $99/month per workspace for unlimited engineers — no per-dev token burn or seat math
- **Free-first LLM routing**: Multi-provider router with OpenRouter, Gemini, Groq, NVIDIA free tiers before paid fallbacks
- **Grounded answers**: Every AI response cites specific files and line numbers from the actual codebase
- **Tenant-scoped derived context**: Temporary clones are removed after indexing; derived documents, embeddings, and graph symbols remain scoped to authorized workspace members
- **SOC 2 Type II audit in progress**: Onramp is not yet certified
- **Controlled deployment roadmap**: Docker artifacts are available; production self-hosting and private-cloud operations are roadmap items

## Supported Languages
Python, JavaScript, TypeScript, Go, Rust, Java — with AST-level parsing for each

## Pricing Summary
- **Free**: 1 repository, 100 AI questions/month, live architecture map
- **Team**: $99/mo (or $82/mo annual) per workspace — unlimited repos, unlimited questions, guided paths, GitHub/Slack/Linear sync
- **Enterprise**: Custom — SSO/SAML, controlled self-hosting, audit logs, dedicated support, and contractual commitments are being defined