Building software, ML infrastructure, and developer tooling out in the open. Our repos start life as public R&D projects, then graduate into commercial platforms once the architecture is proven.
- Learning in public – roadmaps, RFCs, and field notes land in the repos, not private docs.
- Docs-first – every project ships with a microsite, API references, and upgrade guides.
- Operational by default – telemetry hooks, CI templates, and release automation are part of the starter kit.
| Discipline | What we publish |
|---|---|
| Developer Tooling | Prompt ops, background workers, CLI stacks, and SDK ergonomics. |
| ML Infrastructure | Rust/Python accelerators for LangGraph, LiteLLM, CrewAI, and on-device inference. |
| Operational Excellence | Control planes, observability defaults, workflows for release/telemetry hygiene. |
- fast-litellm – Rust accelerators for LiteLLM routing, rate limiting, and pooling.
- fast-crewai – Drop-in shims that speed up CrewAI memory + storage 2-5x.
- fast-langgraph – Cache + checkpoint primitives for LangGraph, written in Rust.
- fastworker – Brokerless task queues plus a control plane/GUI for Python apps.
- savanty – Constraint-solving assistant that turns natural language specs into provable schedules.
- Clone a repo, open an issue, or drop a PR. We review fast and treat community contributions just like internal work.
- Want to commercialize or extend a project with us? Email
hello@neul-labs.com. - Follow along via our website and the “Lab Notes” blog for architecture updates.
Neul Labs Ltd. • Edinbuirgh, United Kingdom