Source: https://internal-agents.com/agents/ramp-inspect # Ramp — Inspect A background coding agent that closes the loop on verifying its own work; runs tests, reviews telemetry, queries feature flags, visually verifies the frontend; now also monitoring production and proposing fixes; also a platform that hosts many internal agents. - Approach type: Background agent - Deployment stage: Scaled - Autonomy: Drafts reviewed - Evidence strength: Mixed - Status: Internal - First reported year: 2026 - Work: Coding, Code review, On-call - Interfaces: Slack, Web, Chrome extension, Github - Invocation: Background, Event driven, Interactive - Entry reviewed: 2026-09-09 ## Where people stay involved - **Inspect coding task → reviewed production merge** — Work product review · Level 3 ## Overview ### Summary A background coding agent that closes the loop on verifying its own work; runs tests, reviews telemetry, queries feature flags, visually verifies the frontend; now also monitoring production and proposing fixes; also a platform that hosts many internal agents. Fact · Reported · High confidence · `ramp-inspect--summary` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) - Supports · [4] [Why We Built Our Own Background Agent (Inspect) (Ramp Builders URL)](https://builders.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-4/content.md) - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · What is Inspect? section ## How it works ### Supporting component Pre-warmed full dev envs; fast cold start; effectively free to run Fact · Reported · High confidence · `ramp-inspect--primitives-0` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Supporting component Server-first agent with a typed SDK + plugin system; code is its own source of truth Fact · Reported · High confidence · `ramp-inspect--primitives-1` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Supporting component Any number of people in one session; each change attributed to its author Fact · Reported · High confidence · `ramp-inspect--primitives-2` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ## Supervision evidence ### Operating model assessment Level 3 for Inspect coding task → reviewed production merge; human attention boundary: work-product-review. Inference · Catalog judgment · High confidence · `ramp-inspect--operating-models-0` Confidence reason: The source explicitly states that Inspect cannot self-approve and production merges retain human review. Qualifications: - Observation date: 2026-05 Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ## Implementation details ### Sandbox Modal sandboxes; per-repo images rebuilt every 30 min from snapshots; warm-on-keystroke; a pool of warm sandboxes Fact · Reported · High confidence · `ramp-inspect--architecture-sandbox` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Harness OpenCode (server-first) as the agent runtime; a plugin blocks writes until sync completes; expanded into production monitoring and self-maintenance; Inspect itself is built with React/Vite, Cloudflare Durable Objects, SQLite, and the Cloudflare Agents SDK Fact · Reported · High confidence · `ramp-inspect--architecture-harness` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Model All frontier models, MCPs, custom tools, skills Fact · Reported · High confidence · `ramp-inspect--architecture-model` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Interfaces slack, web, chrome-extension, github Fact · Reported · High confidence · `ramp-inspect--architecture-interfaces` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Tool access Wired into Sentry, Datadog, LaunchDarkly, Braintrust, GitHub, Slack, Buildkite; monitors production, triages issues, proposes fixes; debugging queries a sanitized read-only production DB replica and Snowflake Fact · Reported · High confidence · `ramp-inspect--architecture-tool-access` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Knowledge Skills that encode how Ramp ships; repo images with the full dev env (Vite, Postgres, Redis, RabbitMQ, Temporal, Chromium, VS Code Server) Fact · Reported · High confidence · `ramp-inspect--architecture-knowledge` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Credentials GitHub auth per user; the sandbox pushes the branch, an API opens the PR with the user's token (no self-approval); production merges retain human review Fact · Reported · High confidence · `ramp-inspect--architecture-credentials` Confidence reason: A linked first-party source states the claim. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ## Reported metrics ### Headline claim 75% of Ramp's merged PRs raised by Inspect sessions (May 2026) Metric · Reported · Medium confidence · `ramp-inspect--headline-metric` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - Reported by: Ramp - Scope: Ramp merged PRs raised by Inspect sessions by May 2026 - Denominator: Merged Ramp pull requests - Observation date: 2026-05 Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 58–68 ### Key observation Around 60% of Ramp PRs authored by Inspect by January 2026 Metric · Reported · Medium confidence · `ramp-inspect--key-metrics-0` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - Reported by: Ramp - Scope: Ramp PRs authored by Inspect by January 2026 - Denominator: Ramp pull requests in the source adoption history - Observation date: 2026-01 Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 58–68 ### Key observation 75% of merged PRs raised by Inspect sessions (May 2026) Metric · Reported · High confidence · `ramp-inspect--key-metrics-1` Confidence reason: A linked first-party source states the claim. Qualifications: - Reported by: Ramp - Scope: Ramp merged PRs raised by Inspect sessions by May 2026 - Denominator: Merged Ramp pull requests - Observation date: 2026-05 Evidence: - Supports · [2] [Ramp x Linear (75% of merged PRs via Inspect)](https://linear.app/customers/ramp) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-2/content.md) · Preserved content.md, lines 10 - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 58–68 ### Key observation Around 30% of merged frontend and backend PRs in the earlier first-party report, after a couple of months of adoption Metric · Reported · High confidence · `ramp-inspect--key-metrics-2` Confidence reason: A linked first-party source states the claim. Qualifications: - Reported by: Ramp - Scope: Merged PRs in Ramp frontend and backend repositories after the first couple of months of adoption - Denominator: Merged PRs in frontend and backend repositories Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) · Preserved content.md, lines 24 ### Key observation Around 90% of PRs merged into the Inspect repository come from Inspect sessions in the later interview report Metric · Reported · Medium confidence · `ramp-inspect--key-metrics-3` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - Reported by: Ramp - Scope: PRs merged into the Inspect repository - Denominator: Merged PRs in the Inspect repository Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 96–102 ### Key observation One million total Inspect sessions crossed in July 2026 Metric · Reported · Medium confidence · `ramp-inspect--key-metrics-4` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - The source does not report the denominator of this figure. - Reported by: Ramp - Scope: Cumulative Inspect sessions crossing the million mark in July 2026 - Observation date: 2026-07 Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 66–72 ### Key observation Under 5 seconds to spin up a fully provisioned remote dev environment Metric · Reported · Medium confidence · `ramp-inspect--key-metrics-5` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - The source does not report the denominator of this figure. - Reported by: Ramp - Scope: Provisioning a fully configured remote development environment Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 100 ### Key observation 150+ engineers contributed to the Inspect codebase in the later interview report Metric · Reported · Medium confidence · `ramp-inspect--key-metrics-7` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - The source does not report the denominator of this figure. - Reported by: Ramp - Scope: Engineers who contributed to the Inspect codebase in the later interview report Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 102 ### Key observation 5.5-person Inspect team (four engineers, a director, and a part-time PM) Metric · Reported · Medium confidence · `ramp-inspect--key-metrics-8` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - The source does not report the denominator of this figure. - Reported by: Ramp - Scope: Inspect team staffing: four engineers, one director, and a part-time PM Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 101 ### Key observation More than 80% of Inspect is written in Inspect sessions Metric · Reported · Medium confidence · `ramp-inspect--key-metrics-9` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - Reported by: Ramp - Scope: Inspect code written in Inspect sessions; distinct from merged PR share - Denominator: Inspect code; the source does not define a code-volume counting method Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 111 ## Reported outcomes and statements ### Key observation Inspect underpins internal agents including ReviewBuddy, Oncall Assistant, Testo, Ramp Research, Voice of the Customer, and error automations Fact · Reported · Medium confidence · `ramp-inspect--key-metrics-6` Confidence reason: A linked participant or independent source reports the claim. Qualifications: - Scope: Examples of internal agents built on Inspect Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Preserved content.md, lines 134–141 ### Key observation Design goal: session speed should be limited only by model-provider time-to-first-token Opinion · Reported · Medium confidence · `ramp-inspect--key-metrics-10` Confidence reason: The source says session speed should only be limited by model time-to-first-token; this is a design goal, not a measured result. Qualifications: - Scope: Target session startup speed, excluding precompleted cloning and installation Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) · Preserved content.md, lines 22 ## Lessons and interpretation ### Lesson Own the tooling; it only has to work on your code, which lets you build something more powerful than off-the-shelf Inference · Catalog judgment · Medium confidence · `ramp-inspect--lessons-learned-0` Confidence reason: The catalog derives this observation from the linked sources. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Lesson Work in public spaces to create virality loops; let the product do the talking, don't mandate Inference · Catalog judgment · Medium confidence · `ramp-inspect--lessons-learned-1` Confidence reason: The catalog derives this observation from the linked sources. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Lesson Ramp argues that a fast background agent can add remote resources and concurrency to the same model Inference · Catalog judgment · Medium confidence · `ramp-inspect--lessons-learned-2` Confidence reason: The catalog derives this observation from the linked sources. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Lesson Move as much as possible into the image-build step so users never wait on setup Inference · Catalog judgment · Medium confidence · `ramp-inspect--lessons-learned-3` Confidence reason: The catalog derives this observation from the linked sources. Evidence: - Supports · [1] [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-1/content.md) ### Lesson The v1 Chrome extension saw little adoption; the pivot to a centrally configured remote dev environment with a coding agent on top drove adoption Inference · Catalog judgment · Medium confidence · `ramp-inspect--lessons-learned-4` Confidence reason: The source reports that v1 saw little adoption and that the November 2025 pivot to a remote dev environment preceded rapid adoption. Evidence: - Supports · [5] [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) · [Preserved copy](https://github.com/steel-experiments/internal-agents-map/blob/main/archive/sources/ramp-inspect-source-5/content.md) · Why build your own background coding agent? section ## Sources 1. [Why We Built Our Own Background Agent (Inspect)](https://engineering.ramp.com/post/why-we-built-our-background-agent) - Engineering blog · First party · Evidence - Original URL: - Accessed: 2026-08-12 · Last verified: 2026-08-31 - Preserved copy in the repository: 2. [Ramp x Linear (75% of merged PRs via Inspect)](https://linear.app/customers/ramp) - Case study · First party · Evidence - Original URL: - Accessed: 2026-08-12 · Last verified: 2026-08-31 - Preserved copy in the repository: 3. [Hacker News project discussion inspired by Ramp Inspect](https://news.ycombinator.com/item?id=48042123) - Hn thread · Community · Discovery - Original URL: - Published: 2026-05 · Accessed: 2026-08-12 · Last verified: 2026-08-31 - Preserved copy in the repository: 4. [Why We Built Our Own Background Agent (Inspect) (Ramp Builders URL)](https://builders.ramp.com/post/why-we-built-our-background-agent) - Engineering blog · First party · Evidence - Original URL: - Accessed: 2026-08-13 · Last verified: 2026-08-31 - Preserved copy in the repository: 5. [Why Ramp built its own in-house coding agent, Inspect](https://newsletter.pragmaticengineer.com/p/why-ramp-built-inspect) - News · Independent secondary · Evidence - Original URL: - Publisher: The Pragmatic Engineer · Published: 2026-08-25 · Accessed: 2026-08-26 · Last verified: 2026-08-31 - Preserved copy in the repository: