Stably AI (Orca)

Seed

We make Orca, MIT open source terminal-based agent orchestrator.

San Francisco, CA, US

Founded

2022

Accelerator

Y Combinator W22

Product Overview

Orca is an open-source, terminal-native agent development environment from Stably AI that orchestrates multiple coding agents in parallel. Released under the MIT license, it lets developers run many AI agents concurrently against the same codebase instead of one at a time.

Each agent works inside its own isolated git worktree, and Orca composes their results through a graph-based orchestration layer that chains, fans out, and reviews agent work. The project is model-agnostic, so teams bring their own API keys and plug in their preferred language models.

Competitive Advantages

Orca's core differentiator is parallelism: it can run multiple coding agents at once, each in a separate git worktree, where most AI coding tools execute a single agent serially. This lets it tackle large changes that one agent cannot finish in a single pass.

Its graph-based orchestrator composes agent output into reviewable runs, the tooling is terminal-native for low overhead, and the entire stack is MIT-licensed and model-agnostic rather than locked to one vendor. Users retain their own API keys, avoiding per-seat SaaS pricing.

Pricing Strategy

Orca is free and open-source software released under the MIT license, so there is no per-seat or subscription license fee. Developers install it locally and run it from the terminal.

Because Orca is model-agnostic and bring-your-own-key, usage cost is driven by whatever AI providers and models the user chooses to connect, rather than by Orca itself. The project's GitHub repository serves as the primary distribution channel.

Product Overview

Competitive Advantages

Pricing Strategy

Orca is free and open-source software released under the MIT license, so there is no per-seat or subscription license fee. Developers install it locally and run it from the terminal.