Engineering memory · MCP-native · Private beta
SenterPoint remembers every fix your team ships — and injects the relevant ones into your AI coding session before you ask. Self-hosted, bring any LLM.
The cost of forgetting
Across every project your team has shipped, the same fix gets re-derived every month — because no one remembers it was already solved in last quarter's tickets.
You start a ticket on a site you haven't touched in six months.
You re-read three folders of .md notes and an old PR thread.
Cost: 90 min
You hit a stack trace that feels familiar.
You Slack-search "that 4xx retry loop" across two workspaces.
Cost: 40 min
You ship a fix that someone else shipped six weeks ago.
You'll discover this next Wednesday when the same ticket reopens.
Cost: +2.5 hours rework
The knowledge already exists in your team's commits — the recall doesn't. For a 10-engineer team that gap is worth ~30 hours a week, roughly $300K a year in fully-loaded engineering time. SenterPoint closes it in <300 ms per session.
Estimate, anchored to Atlassian 2024, McKinsey/IDC, and Stripe/Harris developer-productivity research.
How it works
Nothing to prompt, nothing to file. Precedents show up when your session starts; deeper answers are one question away — and every answer arrives with citations.
You open a project you haven't touched in six months…
When your session starts, SenterPoint detects which project you opened and injects the 5–10 most relevant precedents — past fixes, decisions, gotchas — before you type a word. No prompt, no recall effort.
GET /auto-inject You're about to write a fix…
One question returns the 1–3 precedents that actually apply — with citations pointing at the exact commits and files. The 90 minutes you'd have spent re-deriving: gone.
GET /query · query_knowledge() A precedent drafts itself from your commit metadata. One keystroke to save; nothing hand-written.
POST /precedents Ship a better fix and promote it to canonical. The old one stays searchable, ranks below — no deletes.
POST /precedents/{id}/supersede Ask once; recall spans every project your team owns.
GET /atoms?tags=… The same tools over MCP for your editor, plain JSON over HTTP for scripts.
/openapi.json · /mcp Works where you already write code
Built like infrastructure
Postgres for everything — full-text search, filters, embeddings. No exotic vector store, no orchestration layer, no abstraction tax. The whole service is one process — what runs in your mesh is what runs in our tests.
Private beta · 2 design partners · 50+ sites in production
Single container plus Postgres 16 + pgvector. Bearer-token auth, Caddy reverse proxy, runs anywhere your container does — no managed-service hooks, no upstream phone-home.
your editor ─MCP─▶ senterpoint ──▶ postgres + pgvector ┊ ┊ ┊ └─── your mesh / your VPC / your hardware ───┘
Access granted through private beta invitation. Design partners receive priority onboarding.
Stage C speaks the OpenAI API shape — any provider, any local endpoint (llama.cpp, vLLM, Ollama), any model. Rotate via env var. Your key, your spend, your data path.
Every precedent is scoped to your tenant at the database layer. Isolation is structural — not a WHERE clause we hope to remember.
Built on open source · upstream credit where it's due
Ready when you are
Private beta — tell us about your team and we'll get you set up. Your model, your keys, your hardware.