Brand LogoBrand Logo (Dark)
HomeAI AgentsToolkitsGitHub PicksSubmit AgentBlog

Categories

  • Art Generators
  • Audio Generators
  • Automation Tools
  • Chatbots & AI Agents
  • Code Tools
  • Financial Tools

Categories

  • Large Language Models
  • Marketing Tools
  • No-Code & Low-Code
  • Research & Search
  • Video & Animation
  • Video Editing

GitHub Picks

  • DeerFlow — ByteDance Open-Source SuperAgent Harness

Latest Blogs

  • OpenClaw vs Composer 2 Which AI Assistant Delivers More Value
  • Google AI Studio vs Anthropic Console
  • Stitch 2.0 vs Lovable Which AI Design Tool Wins in 2026
  • Monetizing AI for Solopreneurs and Small Teams in 2026
  • OpenClaw vs MiniMax Which AI Assistant Wins in 2026

Latest Blogs

  • OpenClaw vs KiloClaw Is Self-Hosting Still Better
  • OpenClaw vs Kimi Claw
  • GPT-5.4 vs Gemini 3.1 Pro
  • Farewell to Bloomberg Terminal as Perplexity Computer AI Redefines Finance
  • Best Practices for OpenClaw
LinkStartAI© 2026 LinkstartAI. All rights reserved.
Contact UsAbout
  1. Home
  2. GitHub Picks
  3. ZeroClaw
ZeroClaw logo

ZeroClaw

A single-binary AI agent infrastructure in Rust with swappable providers, channels, tools, and memory backends, built with secure-by-default controls.
15.6kRustMIT
rustai-agentagent-runtimeagent-infrastructureautonomous-agent
tool-calling
sandbox
secure-by-default
single-binary
edge-deployment
raspberry-pi
multi-provider
openai-compatible
telegram-bot
discord-bot
slack-bot
webhook-gateway
memory-system
sqlite
fts5
hybrid-search
vector-search
postgresql
observability
prometheus
alternative-to-openclaw
openclaw-like
self-hosted-automation

What is it?

ZeroClaw turns agent infrastructure into a few‑MB single binary with a trait-driven, swappable core: model providers, messaging channels, tools, memory, tunnels, and observability become config-level choices instead of hard dependencies. Security is the default posture with localhost binding, pairing auth, strict allowlists, workspace scoping, and an optional Docker sandbox for safer execution. Its memory engine favors a local full‑stack approach—hybrid retrieval via SQLite (FTS5 + vectors) and optional PostgreSQL persistence—so you can scale from a terminal assistant to team automation without rewriting the stack.

Pain Points vs Innovation

✕Traditional Pain Points✓Innovative Solutions
Agent runtimes often depend on heavy language runtimes and multi-process stacks, leading to slow cold starts and high memory use on low-spec hosts.A single-binary, trait-driven core makes Provider/Channel/Tool/Memory/Tunnel swappable by configuration, reducing lock-in and glue code.
Security boundaries are frequently bolted on later, making self-hosting risky without extra controls.Security-first defaults: localhost binding, pairing auth, allowlists, workspace scoping, and optional Docker sandboxing for controlled execution.

Architecture Deep Dive

Trait-driven swappable core
Subsystems (Provider, Channel, Tool, Memory, Tunnel, Observer) are modeled as traits with built-in implementations; assembling a runtime becomes a config decision rather than a code integration project.
Localhost-first gateway and security policy
The gateway defaults to 127.0.0.1 and uses pairing codes to mint bearer tokens; inbound channels are controlled by allowlists, and tools are constrained by workspace scoping, path blocks, and command allowlists, with optional Docker sandboxing.
Full-stack memory retrieval (optional Postgres)
SQLite backs keyword search (FTS5/BM25) and vector similarity (BLOB + cosine), merged by weighted ranking for hybrid recall; PostgreSQL can be used as a remote persistence backend when needed.

Deployment Guide

1. One-click bootstrap install (good for fresh or low-spec hosts)

bash
1curl -fsSL https://raw.githubusercontent.com/zeroclaw-labs/zeroclaw/main/scripts/bootstrap.sh | bash

2. Build from source (requires Rust toolchain)

bash
1git clone https://github.com/zeroclaw-labs/zeroclaw.git && cd zeroclaw && cargo build --release --locked

3. Create runtime config and pick a model provider (API key is usually required)

bash
1./target/release/zeroclaw onboard --interactive

4. Run the long-lived runtime or the gateway endpoint (choose as needed)

bash
1./target/release/zeroclaw daemon  # or ./target/release/zeroclaw gateway

Use Cases

Core SceneTarget AudienceSolutionOutcome
Self-hosted multi-channel dev assistant for teamsengineering teams and platform groupsroute CLI/Telegram/Discord/Slack messages into one runtime with consistent tooling and permission boundariesfaster triage and automation with auditable execution
Edge automation agent on low-spec hardwareengineers building store/factory/home automationrun a single-binary agent on Raspberry Pi or cheap instances to execute local commands and report statusstable always-on automation at lower cost
Workflow bot with swappable memory backendssupport and ops teams needing long contextstart with local hybrid SQLite memory and later switch to shared PostgreSQL storagescale memory from personal to team use without rewriting the stack

Limitations & Gotchas

Limitations & Gotchas
  • Most flows require provider API keys or subscription auth; plan secrets management and audit in your platform.
  • Source builds need disk/memory budget; prefer prebuilt releases or bootstrap scripts on low-spec machines.

Frequently Asked Questions

How do I switch model vendors without changing code?▾
Treat the Provider as configuration: pick an OpenAI-compatible or custom endpoint and keep channels/tools the same. The key is stabilizing your prompt contract and tool interface.
How do I scale memory from local SQLite to PostgreSQL?▾
Start with SQLite hybrid retrieval and autosave, then switch the storage provider to PostgreSQL with schema/table config. Use a dry-run migration first to validate indexes and permissions before cutover.
How can I expose the gateway safely?▾
Keep the default localhost bind and use an explicit tunnel such as Tailscale or Cloudflare. Pairing auth, allowlists, and least-privilege tools keep the attack surface bounded.
View on GitHub

Project Metrics

Stars15.6 k
LanguageRust
LicenseMIT
Deploy DifficultyMedium

Table of Contents

  1. 01What is it?
  2. 02Pain Points vs Innovation
  3. 03Architecture Deep Dive
  4. 04Deployment Guide
  5. 05Use Cases
  6. 06Limitations & Gotchas
  7. 07Frequently Asked Questions

Related Projects

DeerFlow — ByteDance Open-Source SuperAgent Harness
DeerFlow — ByteDance Open-Source SuperAgent Harness
26.1 k·Python
Pi Monorepo
Pi Monorepo
14.1 k·TypeScript
QMD
QMD
9.6 k·TypeScript
zvec
zvec
8.2 k·C++