NanoClaw
NanoClaw is an MIT-licensed open-source AI agent runtime maintained by NanoCo that runs every agent group inside an isolated Linux or microVM container and connects to messaging apps via Anthropic's Claude Agent SDK. It targets individual developers, enterprise teams needing defense-in-depth isolation, and government users, and currently generates no direct revenue.
- Company typePrivate
- Founded2026
- HeadquartersNot disclosed
- Headcount1–10
- GTM typeB2B
- OfferingSoftware
What NanoClaw does
NanoClaw is an open-source personal AI agent runtime maintained by NanoCo and released on January 31, 2026 under the MIT license. Architecturally it is a single Node.js host process (roughly 15 source files and ~3,900 lines of code) that orchestrates per-agent-group containers running Anthropic's Claude Agent SDK; each agent group executes inside its own ephemeral Linux container (Docker, Apple Container, or Docker Sandboxes microVM) with explicit mount allowlists, isolated SQLite stores, and FIFO-group queues with concurrency caps. Outbound HTTPS is proxied through OneCLI's Agent Vault so agents never hold raw API keys, and all package installs are routed through JFrog's vetted registries via a June 2026 joint integration. A Skills system ships the trunk registry and infrastructure only; messaging channels (WhatsApp, Telegram, Discord, Slack, Teams, iMessage, Matrix, Google Chat, Webex, Linear, GitHub, WeChat, email) and alternative model providers (OpenAI Codex, OpenRouter, Ollama) are added on demand via Claude-Code-driven /add- skills that copy only the required modules into the user's fork. The V2 release (April 2026) added human-in-the-loop approvals via rich cards and persistent agent-to-agent communication.
The company serves three identifiable segments: technical individual users and developers who want a self-hosted, auditable personal AI assistant across messaging apps; enterprise teams deploying autonomous agents that need defense-in-depth isolation; and government / public-sector users with strict data-sovereignty requirements (Singapore's Minister of Foreign Affairs publicly described NanoClaw as his 'diplomatic second brain'). The Cohen brothers also use NanoClaw to power Qwibit, their AI-first go-to-market agency. GTM is community-led and product-led: a 3-line bash installer, GitHub as the primary distribution hub, Discord for developer relations, multilingual blog, and earned-media coverage in VentureBeat, Fortune, The New Stack, and CNBC.
NanoClaw itself is free; commercial revenue does not yet exist. Underlying Claude API usage is billed by Anthropic with no margin captured by NanoClaw. In May 2026 the company closed a $12M seed round at a $62M valuation led by Valley Capital Partners with participation from Docker, Vercel, monday.com, Slow Ventures, and angel Clem Delangue, while declining a roughly $20M unsolicited acquisition offer. The founders have stated an intent to keep the core open-source and build commercial managed services for enterprises on top, but no enterprise pricing has been publicly disclosed.
NanoClaw firmographics
Firmographics- Name
- NanoClaw
- Legal name
- NanoCo
- Website
- https://nanoclaw.dev
- Company type
- Private
- Founded year
- 2026
- Operating status
- Operating
- Headcount range
- 1–10 employees
- Short description
- NanoClaw is an MIT-licensed open-source AI agent runtime maintained by NanoCo that runs every agent group inside an isolated Linux or microVM container and connects to messaging apps via Anthropic's Claude Agent SDK. It targets individual developers, enterprise teams needing defense-in-depth isolation, and government users, and currently generates no direct revenue.
- Ownership category
- akta.pro rank
Where NanoClaw is headquartered
LocationHeadquarters
- HQ city
- Not disclosed
Offices1 record
Markets served
NanoClaw business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Operations, Marketing or Sales, Infrastructure
Revenue model
- Planned commercial managed services for enterprises: NanoClaw is currently free and open source under the MIT license with no direct revenue stream. The founders plan to keep the core project open-source and build a commercial platform / managed services offering for enterprises on top of it; specifics of the enterprise offering are not yet publicly disclosed.
- Underlying Anthropic Claude API passthrough usage: NanoClaw itself is free; however, it runs on the Claude Agent SDK, which requires a Claude API key or Claude Code subscription. Underlying AI usage is billed by Anthropic (usage-based) — NanoClaw is designed to be lightweight in token consumption but does not collect a margin on this spend.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Freemium | Pay-as-you-go | NanoClaw core — free open source under MIT license |
| Other | — | JFrog security integration — free for open-source developers |
Go-to-market motion2 records
Distribution channels4 records
Marketing channels7 records
NanoClaw product offering
Product offeringCore offering
NanoClaw is a lightweight, open-source personal AI agent runtime released under the MIT license that runs on a user's local machine (macOS, Linux, or Windows/WSL2). It orchestrates per-agent-group containers running Anthropic's Claude Agent SDK, connects to messaging platforms such as WhatsApp, Telegram, Discord, Slack, Microsoft Teams, iMessage, and others, and routes all outbound HTTPS requests through OneCLI's Agent Vault so agents never hold raw API keys. The maintainers plan to keep the core project open-source while building commercial managed services for enterprises on top of it.
Product overview
NanoClaw is a single, unified open-source product (not a multi-product suite): one Node.js host process that orchestrates per-session agent containers running the Claude Agent SDK. It is built around two architectural pillars — (1) the core NanoClaw runtime (with the current generation tracked as NanoClaw V2, which added human-in-the-loop approvals and persistent agent-to-agent communication), and (2) the Skills System, the on-demand extensibility layer that replaces traditional feature bundles. Messaging channels (WhatsApp, Telegram, Slack, Discord, Teams, iMessage, Matrix, Google Chat, Webex, WeChat, Linear, GitHub, email via Resend), alternative model providers (OpenAI Codex, OpenRouter, Ollama), container runtimes (Docker, Apple Container, Docker Sandboxes), and credential/security layers (OneCLI Agent Vault, JFrog) are all delivered as Skills or partner integrations rather than as separate NanoClaw products. The maintainer entity NanoCo (raised $12M seed at a $62M valuation in May 2026) is building commercial managed services for enterprises on top of the open-source core, but no separate managed product is currently named in the sources.
Differentiator
Problem solved
Functional benefit
Products and services
- NanoClaw Lightweight, open-source personal AI agent runtime (MIT licensed) that runs every agent group inside isolated Linux/Apple containers, connects to messaging apps such as WhatsApp and Telegram, routes outbound credentials through OneCLI's Agent Vault, and is built directly on Anthropic's Claude Agent SDK. Designed as a single Node.js process with ~15 source files and ~3,900 lines of code for full auditability in approximately 8 minutes.
- NanoClaw V2 April 2026 major version release of NanoClaw that adds human-in-the-loop approvals via rich cards inside messaging apps, persistent agent-to-agent communication, and Flexible Isolation V2 supporting per-channel agent wiring, shared sessions across channels, and agent swarms.
Quantifiable outcome
- 7,000+ GitHub stars in first week; 30k+ GitHub stars total
- +2 more outcomes
Companies that use NanoClaw
Customer profileNamed customers2 records
Segments3 records
Ideal customer profiles3 records
NanoClaw technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
Integration23 records
AI capability9 records
Feature7 records
NanoClaw partnerships and signals
Strategic signalPartnerships
Six partnerships are on record, tiered core, minor and flagship.
- JFrogcoreJoint 'immune system' security integration that protects NanoClaw agents from downloading malicious code by routing all package requests exclusively through JFrog's vetted software registries. Blocks compromised packages and auto-guides agents to safe versions. Free for open-source developers; integrates with enterprise JFrog environments.
- exe.devminorNanoClaw published a step-by-step guide to deploying the agent runtime on exe.dev, where users sign up, activate a subscription, spin up a VM, and run the NanoClaw setup.
- OneCLIcoreOneCLI's Agent Vault acts as a credential-injection gateway for every NanoClaw agent, so agents never hold raw API keys. The Vault enforces per-agent policies, rate limits, and approvals on outbound HTTPS requests.
- DockerflagshipNanoClaw integrates with Docker Sandboxes so each AI agent runs in its own isolated container inside a microVM rather than directly on the host. Docker also participated in NanoClaw's $12M seed round alongside Vercel, monday.com, and Slow Ventures.
- Anthropic (Claude Agent SDK)coreNanoClaw natively runs Claude Code via the official Claude Agent SDK. Underlying AI usage is billed by Anthropic; Singapore has signed broader agreements with Anthropic and OpenAI as referenced in related coverage.
- NanoCocoreNanoCo is the maintainer entity behind NanoClaw; the homepage banner reads 'Meet NanoCo, maintainers of NanoClaw: we raised $12M to give every member of your team a professional assistant' with a link to nanoco.ai.
Scale indicators5 records
Recent moves6 records
Expansion highlights6 records
NanoClaw competitors and assessment
Company assessmentBroad incumbents
- LangChain: LangChain is the broadest open-source LLM-application framework, including its LangGraph agent-orchestration layer. Overlaps with NanoClaw on agent workflows but is a much larger, broader-incumbent platform rather than a security-focused personal-agent runtime.
- AutoGen (Microsoft Research): AutoGen is Microsoft's open-source framework for building multi-agent AI applications with conversable agents. Comparable in agent-orchestration focus but backed by a large incumbent (Microsoft) with broader enterprise distribution than NanoClaw.
Emerging players
- n8n: n8n is an open-source workflow automation platform with growing AI-agent capabilities and self-hosting options. Overlaps with NanoClaw on local execution and automation of multi-step tasks, though targeted at workflow builders rather than messaging-driven personal agents.
- Botpress: Botpress is a conversational-AI platform supporting many of the same messaging channels (WhatsApp, Telegram, Slack, Teams) that NanoClaw targets. Comparable on multi-channel agent deployment, but oriented toward enterprise chatbot builders rather than individual-user local agents.
Others
- Ollama: Ollama is the leading open-source local model runtime, integrated into NanoClaw via the /add-ollama-provider skill. Not a direct competitor but a complementary infrastructure layer that NanoClaw users frequently combine with the runtime.
- Hugging Face: Hugging Face is the dominant open-source AI model and application hub and an investor ecosystem for open-source AI (Clem Delangue, its CEO, angel-invested in NanoClaw). Ecosystem peer rather than direct competitor, but relevant for distribution and developer-mindshare benchmarks.
Direct peers
- Letta: Letta (formerly MemGPT) is an open-source framework for building stateful, deployable AI agents with memory and tool use. Comparable to NanoClaw as an open-source agent runtime targeting developers and technical users, though focused more on hosted memory than on container-isolated security.
- CrewAI: CrewAI is an open-source multi-agent orchestration framework that composes role-based AI agents to complete tasks. Closely comparable to NanoClaw's agent-group model and V2 agent-to-agent communication, with overlapping target user (technical builders).
- OpenClaw: OpenClaw is the open-source AI agent framework NanoClaw explicitly positions against on its homepage, framed as the monolithic (~434k lines) alternative that NanoClaw contrasts with its ~3,900-line security-first runtime. Most directly comparable peer.
- Open Interpreter: Open Interpreter is an open-source project that lets language models execute code locally on a user's machine. Directly comparable to NanoClaw on the local-machine, terminal-driven, self-hosted personal AI agent thesis, though without NanoClaw's messaging-channel and container-isolation focus.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks6 records
Key highlights7 records
Customer concentration
NanoClaw social profiles
Digital presenceNanoClaw financial estimates
Financial estimateRevenue estimate
Valuation estimate
NanoClaw leadership team
Management profileNumber of profiles
Profiles1 record
NanoClaw subsidiaries and ownership
Company hierarchySubsidiaries1 record
NanoClaw funding detail
Funding detailFunding overview
Funding rounds
Investors
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
NanoClaw M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about NanoClaw
What does NanoClaw do?
NanoClaw is a lightweight, open-source personal AI agent runtime released under the MIT license that runs on a user's local machine (macOS, Linux, or Windows/WSL2). It orchestrates per-agent-group containers running Anthropic's Claude Agent SDK, connects to messaging platforms such as WhatsApp, Telegram, Discord, Slack, Microsoft Teams, iMessage, and others, and routes all outbound HTTPS requests through OneCLI's Agent Vault so agents never hold raw API keys. The maintainers plan to keep the core project open-source while building commercial managed services for enterprises on top of it.
Is NanoClaw a public or private company?
NanoClaw is a private company. It is classified as venture growth investor backed and is currently operating.
When was NanoClaw founded?
NanoClaw was founded in 2026. It employs 1 to 10 people.
Where is NanoClaw based?
NanoClaw is headquartered in Not disclosed.
How does NanoClaw make money?
Two revenue lines are on record. Planned commercial managed services for enterprises are the primary driver. The others are underlying Anthropic Claude API passthrough usage.
Who are NanoClaw's main competitors?
Broad incumbents on record are LangChain and AutoGen (Microsoft Research). Emerging players are n8n and Botpress. Others are Ollama and Hugging Face. Direct peers are Letta, CrewAI, OpenClaw and Open Interpreter.
Does NanoClaw have an API?
No public API is recorded for NanoClaw.