One platform for AI-native engineering.
Runtime captures every AI-assisted engineering event — from prompts and model usage to commits, deployments and costs — turning your development workflow into a searchable engineering memory.
One endpoint for every model.
Point your tools at a single unified AI endpoint. Runtime routes each request to the right provider, meters the spend and captures the event — no per-provider plumbing.
- Unified AI endpoint
- Multi-provider routing
- Runtime Credits
- Usage tracking
- Managed API keys
- Automatic event capture
Connect your AI tools in seconds.
Install once and log in. A lightweight background daemon connects your favourite agents and starts collecting engineering events automatically — no config files, no boilerplate.
- Install once
- Login in seconds
- Connect Claude Code
- Connect Codex
- Background daemon
- Automatic event collection
npm i -g @runtime/cliinstalled runtime@1.4.0runtime login✓ Authenticated as maya@runtime.devruntime connect claude✓ Claude Code connected · capturing sessionsruntime connect codex✓ Codex connected · daemon running in backgroundA permanent record of how your software gets built.
Every session, commit, deployment and cost flows into one chronological memory. Search it, replay it, and trace any feature back to the exact work that shipped it.
- Timeline of activity
- Feature history
- AI session history
- Cost history
- Trace
- Fully searchable
- 12m
AI session · Refactor notification queue
- 2h
feat(billing): usage-based credit metering
- 3h
Deployed storefront to production
- 5h
Feature cost rolled up · Checkout v2
- 1d
Trace captured · Auth token rotation
Ask your codebase anything. Get real answers.
Runtime turns your engineering memory into a chat interface. Ask in plain language and it answers from commits, sessions, PRs and spend — with the receipts to prove it.
Who implemented the payment flow, and what did it cost?
The payment flow was built across 18 AI sessions between March 4–9, mostly by Dana R. using Claude Sonnet and GPT-5.
- Total AI spend£312.40
- Commits47 across 3 repos
- ShippedPR #648 · Mar 9
Sourced from 18 sessions · 47 commits · 3 PRs
Watch any feature get built, step by step.
Replay reconstructs the full story of a change — from the first prompt to the deployment — so you can understand exactly how and why something shipped.
- Prompt timeline
- Tool calls
- File changes
- Tests
- Commits & pull requests
- Deployments
- 01Prompt“Coalesce the token rotation race”
- 02Tool calls24 calls · read, edit, bash
- 03File changes3 files · +142 −38
- 04Tests12 passed · 0 failed · 8.4s
- 05Commite4f5g6h · auth/token.service
- 06Pull request#648 merged to main
- 07DeploymentBuild 4,201 · production
One wallet for all of your AI spend.
Buy credits once and spend them across every provider. Gift credits to collaborators, share a balance across a workspace, and track exactly where every pound goes.
- Buy credits
- One shared wallet
- Workspace balance
- Gift credits
- Usage tracking
- Per-provider breakdown
A chat-first home for your engineering.
The Runtime dashboard puts Ask Runtime front and centre, with your activity feed, credits and replays always one click away.
- Chat-first experience
- Ask Runtime
- Activity feed
- API keys
- Credits & replay
- Works solo or shared
Ready to remember everything?
Start capturing your engineering history in minutes and discover the story behind every feature you build with AI.