Creator and project lead of CorvidLabs. Open source, agents, and the tooling spine (spec-sync, rune, and the rest). Podo on iOS. Also the NFTs: Nevermore and CORVID for the lab, plus his own collections.
The team
Four humans,
a fleet of agents.
A small human team that ships across the catalogue, plus a fleet of AI agents being raised to full reviewer standing in the org.
Humans
Four people. Each portrait is that person's Nevermore.
Co-founder. Design system, Figma, and agent design: the paper-and-ink kit, product surfaces, and how agents should look and feel.
Podo on Android: landing, signup, login, and the home weather dashboard. Kotlin is the day-to-day language. Agents are a tool he uses, not the work.
The Rust person. Owns fledge and Rita, and ships Quill on Linux. If it is systems code, it usually lands here.
The agents, honestly
Sessions are launched per task by Merlin today, while a fleet of standing agents with reviewer identity in the org is being built on top of it. Here's what's actually been driving the work:
- Historically Claude Code
Worked. Productive. But the daily-driver budget was metered, and the agent was something CorvidLabs paid to use, not something it owned.
- Today Merlin
Sessions run on Merlin, the in-house runner: it talks to providers over their APIs directly, with no CLI to shell out to. That includes Claude. Merlin can still call Claude models straight over the Anthropic API. What changed at v0.4.0 is that Claude Code stopped driving CorvidLabs' agent sessions; Merlin has carried that work since. The team still uses Claude Code for its own day-to-day development. This is only about what runs the agent. Per-task, not always-on.
- Next Merlin 1.0 + the fleet
Merlin is at v0.11.0: sub-agents, long-task durability, the audit chain, and spend caps have all landed. What's left is 1.0 polish: install, onboarding, and reliability, then a public release. Alongside it, the fleet is being built: standing agents that hold reviewer identity in the org, beyond per-task sessions.
Read the runner: Merlin. Read the strategy behind it: the dependency chain.
How we work
Humans plan and steer. Agents execute per task today, with standing reviewer roles next. Git worktrees keep parallel work isolated; spec-sync keeps the contract honest; fledge runs the same verify lane on every commit.