An invite-only community whose own platform is rebuilt by an AI harness off GitHub issues — and every member gets the full source. Watch each change land on a live Kanban board, or open a feature request and pressure-test the loop yourself.
Slash commands in chat
Member directory
Dark mode for the feed
Realtime presence dots
Export kanban to CSV
Tooling, clouds & platforms covered
Membership opens up the platform itself — its source, its AI build harness, and a live board of everything changing — alongside working cloud tooling you can use in production.
Forget Slack and Discord. Every member gets the platform's full source — and you watch it change in realtime on a live Kanban board: issues opening, PRs merging, deploys shipping, as they happen.
Open a feature request to bend the platform to your will. Each one becomes a GitHub issue an autonomous AI harness picks up — you're stress-testing a live build loop running in production.
Hands-on access to the harness itself: an OpenAI Symphony-style orchestrator that turns GitHub issues into shipped code. Open an issue, watch it plan, build, review, and merge — end to end.
Every member gets the platform's entire source — frontend, backend, infra, and the harness. Clone it, run it locally, send a PR. The thing you're a member of is the thing you can change.
Connect and manage multiple clouds with one consistent developer workflow — accounts, regions, identity, and resources behind a single command surface.
Direct access to the engineers building the platform — questions, design reviews, and the running commentary behind every change on the board.
Across the community and the resources, we cover the full stack of modern platform engineering — from cloud accounts to AI tooling.
Members get access to the full set — these are condensed previews of what each repo looks like.
The platform you're a member of — chat, feed, and a live Kanban board. Yours to read, run, and reshape.
$ innerloop board watch --live► card #418 "dark mode for the feed" todo → in-progress► pr #1204 feat: realtime presence dots ✔ merged · deploy queued► card #421 "export kanban to CSV" backlog → todo↻ streaming events… ⌃C to exitOpenAI Symphony-style build harness — turns GitHub issues into planned, reviewed, and merged pull requests, autonomously.
$ harness run --issue 421✔ planned 3 steps · 2 files✔ wrote lib/export.ts (+74)✔ tests 6 passed✔ opened PR #1207 → review⏳ awaiting maintainer approvalOne CLI for AWS, Azure and GCP — consistent identity, accounts, and resource ops.
$ innerloop login --provider aws --account prod✔ assumed role arn:aws:iam::4719…:role/PlatformAdmin$ innerloop k8s ctx --cluster eks-prod-use1✔ kubeconfig merged · 14 nodes · 312 pods$ innerloop get deployments --owner platformNAME ENV REGION OWNERgateway-api prod us-east-1 platformidentity-svc prod us-east-1 platformnotifications prod eu-west-1 platformBackstage app + plugins that ingest your AWS/Azure/GCP/Cloudflare estate into a single catalog.
apiVersion: backstage.io/v1alpha1kind: Componentmetadata: name: gateway-api annotations: aws.com/account-id: "471900000000" cloudflare.com/zone: edge.example.comspec: type: service lifecycle: production owner: team-platform system: ingressPrototype 'company brain' — connects infra, ownership, docs, and AI search across your platform.
const answer = await brain.ask({ q: "who owns gateway-api in prod?", scope: ["catalog", "runbooks", "oncall"],});// → team-platform · #plat-oncall · runbooks/gateway-api.mdLightweight AWS cost & usage insights — daily summaries, anomaly nudges, per-team views.
$ babyctl cost summary --window 7dSERVICE 7d ($) Δ vs prev noteEC2 4,128 +6.1% new c7g fleetS3 612 -2.3% —NAT Gateway 1,907 +18.4% ⚠ anomaly · vpc-prod-1Paid keeps the bar high. It filters for engineers who actually want to build platforms, not collect bookmarks.
~ a sketch of the kind of system we ship recipes for
No newsletter recycling. No engagement farming. Resources are written by engineers shipping platforms, for engineers shipping platforms.
Working code, not architecture diagrams. Every recipe maps to a real production constraint — IAM, cost, latency, lifecycle.
Drop-in CLIs, Backstage plugins, Actions, and Terraform modules — designed to be forked and adapted, not just admired.
We'd rather ship five solid repos than fifty stubs. Everything in the org has been used to solve a real problem.
No takes. No threads. No conference fluff. Just patterns, code, and discussion among practitioners.
Ask the people building these systems. Implementation reviews, design feedback, and post-mortems — in private.
Pay once, keep access. No recurring billing. No auto-renewal. No seat upsells.
If something isn't covered here, email us — we'd rather answer directly than fluff out a help center.