Skip to content

Neul Labs is not Neuland Laboratories

Neul Labs (this company, Neul Labs Limited, UK Companies House #SC767862) is not Neuland Laboratories Ltd. (an Indian pharmaceutical manufacturer, NSE: NEUL). The two companies have no connection. If you are looking for the Indian pharmaceutical company, see neulandlab.com. If you are looking for the UK open-source AI infrastructure studio (this site), you are in the right place.

Developer infrastructure for the agentic era

Neul Labs is a UK-based open-source studio. We ship the boring-but-critical pieces that the agent stack is missing — isolated workspaces, durable memory, safe data access, payments, deployment, and Rust-accelerated drop-ins for the Python libraries agents already depend on. The project catalogue links to the source and documentation surfaces that can be inspected for current availability, licence and version.

We also provide scoped engineering services and work with technology providers through evidence-based partnerships. Provider names describe compatibility or a proposed integration unless an awarded status is stated explicitly.

Our Thesis

The industry is moving from chatbots to autonomous agents. Agents need an actual operating environment — isolated workspaces, durable memory, safe data access, payment rails, orchestration, and deployment — not just a chat completion endpoint.

Most of that layer doesn't exist yet. The pieces that do exist (LiteLLM, LangGraph, CrewAI) were built for prototypes, not production: connection pooling, checkpointing, and serialization all become bottlenecks under real load.

We build both halves. New primitives where the stack is missing them (agentvfs, brat, fastagentic, ormai, mcp-pay, memorg). Rust-backed drop-ins where the Python libraries already won developer adoption but can't carry production weight (fast-litellm, fast-langgraph, fast-crewai, fast-axolotl).

Our Approach

1

Find the missing primitive

Either a hot path in an existing Python library (LiteLLM, LangGraph) or a piece of agent runtime that nobody has built yet (workspaces, payment-aware MCP, policy-enforced ORMs).

2

Build it in Rust where it matters

Rust + PyO3 for accelerators. Rust for daemons, runtimes and append-only event logs. Go for the workflow plane. Use the right tool per layer.

3

Ship as drop-in

If we can preserve the existing CLI/API surface (jest, pytest, ninja, wget, curl, litellm, langgraph) we do. One import or one binary swap, no config.

4

Make the work inspectable

Publish source, tests, methods and limitations where the project permits it. Licence and release status are stated per repository rather than inferred across the portfolio.

Open Source First

Public source makes engineering choices inspectable and lets another team test whether a component fits. It does not replace project-specific licence, security, release, maintenance and support review.

Discoverable source

GitHub and package registries can expose useful work when the metadata, licence and release artefact are accurate.

Visible feedback

Public issues and releases can show how maintainers respond to defects and compatibility changes; an empty issue tracker proves nothing.

Inspectability

Reviewers can inspect implementation and tests, while still performing their own dependency, security and deployment assessment.

Composable boundaries

Small tools can be evaluated independently and composed through documented interfaces rather than adopted as one platform.

Contribution path

A useful contribution guide, security route and maintainer response make external participation possible; adoption is never assumed.

Service boundary

Scoped engineering, support or managed work is contracted separately from the licence and support status of a public repository.

Technology Stack

Rust

Core performance engine. Memory safety, zero-cost abstractions, fearless concurrency.

Python

PyO3 integration layer. Seamless interop with the AI ecosystem.

Go

Workflow automation. Single-binary deployments, native concurrency.

PyO3

The bridge. Zero-overhead Rust-Python bindings for drop-in compatibility.

For Investors

Where We Play

Developer infrastructure for the agentic era — picks and shovels, not the next chatbot. Adjacent to LiteLLM, LangGraph, CrewAI, MCP, Claude Code, Aider and the rest of the agent ecosystem; complementary rather than competitive.

Why Us

Rust + Python interop at scale, agent runtime design, and shipping discipline. The portfolio already covers eight categories — accelerators, agent infrastructure, runtimes/workspaces, dev tools, build/release, network/data, compliance, and personal AI — with the same opinionated stack underneath.

Business Model

Public engineering work provides inspectable capability evidence. Current revenue paths include scoped systems, performance, integration and developer-tooling engagements; any support or managed service is defined by a separate written scope rather than implied by a repository licence.

Traction

The website maintains a project catalogue across the agent and developer-tooling stack. Source and some release artefacts are published through GitHub and language package ecosystems; availability is verified per project. See the projects page for the index, or browse the current organisation repositories on GitHub.

Get in Touch

Interested in engineering support, partnering, investing, or contributing? We'd love to hear from you.