Gergely Orosz · 8 min read · Gen AI Tools
Why Ramp built its own in-house coding agent, Inspect
Read the original on newsletter.pragmaticengineer.com →
Ramp built Inspect, an internal remote-sandbox AI coding agent, to run multiple agents in parallel, access Ramp’s internal data sources, and verify changes end-to-end for both backend and frontend. It differs from public coding assistants by offering centralized tooling, rapid sandbox spin-up, and deep internal integrations, which have driven strong adoption. More than 150 engineers have contributed to the project, and 75% of merged PRs are now raised by Inspect.
Key Takeaways
- Internal remote-sandbox AI coding agent enables parallel sessions, centralized configuration, and broader collaboration.
- Deep internal integrations and access to Ramp’s own tools differentiate Inspect from off-the-shelf AI coding assistants.
- Built-in verification loop checks backend tests, telemetry, feature flags, and frontend visuals to ensure changes work.
- Open collaboration culture with public sessions and wide contributor base (over 150 engineers).
- High adoption reflected in 75% of merged PRs being raised by Inspect.
Excerpt
At a select few tech companies, they write most of their code with their own, custom-built, internal AI coding agents. This is different from most of the industry which uses AI coding agents and harnesses like Codex, Claude Code, Cursor, OpenCode, GitHub Copilot, etc. At Ramp, their own version is called Inspect , while at Block it’s Goose (open source), at Stripe it’s Minions , and River at Shopify. But why not just use what frontier labs and coding harness AI startups already offer; why take the time and effort? We reached out to Ramp, a fintech company big on…