Live, hand-built dashboards over Lawhive's engineering activity. Unlike the generated code-analysis charts, these are their own page and recompute from live data on view.
Every tRPC procedure and HTTP endpoint across the platform-v3 API surfaces, resolved from the code: where each lives, who can call it, whether it really reads or writes (vs. what its name claims), how many service hops it makes, which Inngest events it triggers, and whether anything still calls it. Links to the domain model and Inngest dashboards.
A living registry of what the platform is made of: twelve business domains and every model across 16 services, with an agent-written interpretation layer — what each domain and entity means, how they relate, and the flows between them. Searchable, regenerable, linked to source.
The same treatment for the monolith: 17 business domains and 146 models across three datastores, with an agent-written interpretation layer. Because v2 has no service boundaries to read ownership from, extraction also scans every Prisma and Drizzle call site, so each table shows which module writes it, what reaches it only through a nested write, and what nothing references at all.
Every Inngest event, function, and cross-repo chain across platform-v3, platform-v2, and ai-platform, extracted from source (TypeScript + Python). Browse events, functions, and the chains between them, with proposed naming normalization inline.
A review economy: everyone starts with 20 coins, and reviewing a PR or tech plan earns a share of its bounty from the author. Leaderboard, open bounties, and a live coin feed across the production repos.