One console for the whole substrate.
Feed it, ask it, query it, watch it, govern it — lakecode's console is where your engineers and your agents work the same substrate. Every view is grounded, cited, and yours to inspect. Here's each surface in full.
Prefer email? Reach the founder directly.
Every action your team takes flows through Cowork.
Each engineer's session, each agent's run, every proposed claim and review — across the org, in one live feed. It's how your team and their agents stay in sync, and nothing becomes org truth until a human approves it.
Webhook retries need idempotency keys
How does the refund reconciliation job avoid double-counting?
approved “Stripe client must use the async transport”
Analyzer pinned to temp=0 → retrieval is bit-stable
databricks-sdk
rejected “Use a global mutable cache for tokens”
Governance you can watch: claims arrive proposed, a reviewer accepts or rejects, and only then do they compound into what the next session — anyone's — is grounded on.
Ask the expert. Get a cited answer.
Ask a service, a topic, or your whole org a question in plain English. lakecode compiles the relevant claims into a grounded answer with inline citations you can click — scoped expertise on tap, for your engineers (just ask) and your other agents (over MCP).
youWhy does checkout 500 during retry storms?
Retries without idempotency keys replay the charge — that's the 500. The client already enforces idempotent-key-only retriesc1, and the gateway requires the outbound header from INC-4471c2. Apply both and the storm stops double-charging.
a service
The payments expert.
Why retries are idempotent-key only, which IDs are valid, the gateway allow-list from INC-4471 — everything ~/payments-api has taught it.
a topic
The lakehouse expert.
How warehouse.sessions derives, why timestamps are UTC end-to-end, what locked prod in March — the data domain, across every repo that touches it.
your whole org
The org expert.
Conventions, decisions, and failed paths across every team — promoted local → workspace → org and reviewed, so org truth is earned, not assumed.
Not a wiki-bot guessing from documents: every answer is grounded in cited claims with provenance, and it sharpens every session.
Point it at your sources. The substrate fills itself.
Repos, incidents, docs, Databricks jobs, chat — lakecode reads each on a schedule, extracts the claims and entities buried in them, and writes them to the substrate with provenance intact. The more it ingests, the sharper every expert downstream gets.
Every claim remembers where it came from. The moment it grounds an answer, that answer is citable straight back to the source — repo line, incident, or job run.
It's just data. Query it like it is.
The substrate is warehouse-native: entities and claims in tables you can browse, join, and audit. Plain SQL over everything it knows — with confidence and provenance as columns, not a black box you have to trust.
select e.name, c.claim, c.confidence, c.source
from entities e join claims c on c.entity_id = e.id
where e.corpus = 'payments-api' and c.reviewed
order by c.confidence desc limit 5;No proprietary lock-in: it lives next to your data in your warehouse, and you can read it directly — the same store the experts and agents compile their context from.
The gate between proposed and trusted.
Cowork shows you everything; the review queue is where you act on it. Each pending claim opens with its evidence, its confidence, and the exact sessions it would ground — accept it, reject it, or send it back. Only what a human accepts becomes org truth.
Webhook retries need idempotency keys
proposed by bob · session #412 · payments-api · 6m ago
If accepted, grounds 3 open sessions and supersedes claim 4298a000 · “retries are safe to replay”
Promotion is earned, not assumed: local → workspace → org, each step a sign-off. That discipline is exactly why an agent can trust what it reads.
Manage every agent your org has deployed.
One place for the whole fleet: which agents run where, what version each is on, who holds keys, and what every one has written back — rolled out and governed together.
| agent | scope | version | status | governance |
|---|---|---|---|---|
payments-bot | ~/payments-api | v2.4.1 | active · 3m ago | 12 claims pending review |
lakehouse-etl | warehouse | v2.4.1 | active · 1h ago | clean |
mobile-ci | ~/mobile | v2.3.0 | update available | 3 claims pending review |
support-bot | ~/support | v2.4.1 | paused | clean |
Roll out a version, rotate a key, or pause an agent across the org — and every claim an agent wrote traces back to the session that produced it.
See it on your own codebase.
The console runs on your substrate — ingested from your repos, grounded in your claims. During the private beta, we set it up with you.