HybridClaw

Business Skills That Work On Smaller Models

HybridClaw's main promise is practical business work. The bundled CRM, marketing, analytics, finance, cloud, office, and operations skills are implemented as tested helpers with eval scenarios, credential boundaries, and approval tiers, and are validated against Qwen/Qwen3.6-27B-FP8 as the small-model baseline.

That is the useful difference: a compact model can operate Salesforce, HubSpot, GA4, Google Ads, Lexware, Airtable, Miro, Zoho, Microsoft 365, warehouse SQL, invoices, cloud ops, and office documents through structured rails instead of fragile free-form prompting.

HybridClaw also treats agents as networked coworkers. Local agents, hosted HybridAI proxy agents, and trusted peer HybridClaw instances can address one another, exchange end-to-end encrypted A2A envelopes, and route work through approval-aware channels. An A2A local mode keeps peer delivery and authenticated administration reachable while disabling other external gateway and channel surfaces.

First-run onboarding is built around hatching: a new agent asks about the user's work, records useful context, keeps setup links visible in chat, and can send a tailored first-jobs welcome email when an email route is available. The bootstrap lifecycle is recorded as structured onboarding audit events, including messages, welcome mail, workspace file updates, completion, and abort paths.

Credentials stay outside the model context. Secrets live in the encrypted runtime store and SecretRef-backed tools resolve them at execution time, so LLMs see the requested action and approval context, not raw API keys, passwords, or bearer tokens.

Run one local assistant, operate a fleet of role-specific coworkers, or launch HybridClaw on HybridAI Cloud in a few minutes at hybridclaw.io.

Quick Start · Launch Cloud · Installation · Docs · Configuration · Commands · Contributing

Why HybridClaw

| You need | HybridClaw gives you | | --- | --- | | A first run that becomes useful quickly | Guided hatching with setup links, tailored first-job suggestions, optional onboarding-specific model routing, welcome-email handoff, and structured audit events | | Business workflows that survive real use | Production skill helpers with fixtures, eval scenarios, targeted tests, approval tiers, and a Qwen/Qwen3.6-27B-FP8 validation baseline | | Generated work artifacts you can reuse | An Apps gallery for self-contained HTML apps, dashboards, documents, games, tools, live connector-backed views, sharing links, and Teams tabs | | Multi-agent workflows across installations | Local agents, hosted proxy agents, encrypted A2A trust, explicit addressing, inbound envelopes, reply-back delivery, admin-visible peer pairing, and an A2A-only deployment mode | | Credentials the model cannot read | Encrypted runtime secrets, SecretRef-backed execution paths, and scoped gateway API tokens that keep raw keys and passwords out of prompts and tool results | | Assistants that can act, not just chat | A gateway, web chat, Apps gallery, TUI, admin console, scheduler, tools, and OpenAI-compatible API behind one local service | | Control over sensitive work | Approval policy, sandbox boundaries, output guardrails, and hash-chained audit trails | | Agents that fit existing teams | Discord, Slack, Teams, Telegram, email, voice, web, and more through the same runtime, with WhatsApp available as an install-on-demand plugin | | Operational memory | Local files, SQLite state, semantic recall, session compaction, and optional HybridAI cloud memory | | Repeatable expert workflows | Per-agent workspaces, budgets, deterministic model-tier routing, A2A trust, proxy agents, .claw archives, and human-distillation workflows |

Install

Fastest managed launch: HybridClaw on HybridAI Cloud.

Apple Desktop App for macOS:

• Download the signed and notarized Apple Silicon DMG from GitHub Releases.
• Open the DMG, drag HybridClaw.app into /Applications, and launch it.
• The desktop app starts the local gateway and opens the chat, agents, and admin surfaces in a native macOS window.

Linux/macOS CLI one-line installer:

curl -fsSL https://raw.githubusercontent.com/HybridAIOne/hybridclaw/main/scripts/install.sh | bash

Manual npm install:

npm install -g @hybridaione/hybridclaw
hybridclaw onboarding
hybridclaw gateway
hybridclaw tui

Requirements: Node.js 22. Docker is recommended for the default sandbox.

First Run

After the gateway starts, open:

| Surface | URL / command | Use it for | | --- | --- | --- | | Web Chat | http://127.0.0.1:9090/chat | Chat, slash commands, model and agent switching | | Apps Gallery | http://127.0.0.1:9090/apps | Generated web apps, documents, games, tools, live connector-backed views, and sharing | | Admin Console | http://127.0.0.1:9090/admin | Searchable settings plus agents, automation, activity, connectivity, security, and federation workflows | | Agents UI | http://127.0.0.1:9090/agents | Agent fleet overview and prompt-file editing | | TUI | hybridclaw tui | Terminal chat, approvals, status, resume | | OpenAI-compatible API | http://127.0.0.1:9090/v1/chat/completions | Local evals and compatible clients |

For signed macOS desktop builds and future architectures, use the GitHub Releases page.

Desktop wrapper from source:

npm install
npm run desktop

What You Get

| Area | Built in | | --- | --- | | Skills | 79 bundled skills, production business helpers, eval fixtures, packaged skill lifecycle, and human-distillation workflows | | Channels | Discord, Slack, Signal, LINE self-chat, Telegram, Microsoft Teams, email, iMessage, fax, Twilio voice, web, incoming webhooks, and an install-on-demand WhatsApp plugin | | Runtime | Gateway service, TUI client, web chat, Apps gallery, searchable admin console, loopback OpenAI-compatible API, and Docker or host execution with document and spreadsheet tooling | | Governance | Encrypted runtime secrets, scoped API tokens, SecretRef credential isolation, approval policies, sandbox controls, hash-chained audit trails, dependency license gates, SBOMs, and third-party notices | | Memory | Local memory files, SQLite persistence, semantic recall, session compaction, optional HybridAI cloud memory sync | | Agents | Guided hatching, per-agent workspaces, models, budgets, prompt files, deterministic tier routing, explicit addressing, proxy agents, encrypted A2A trust, peer-instance communication, and reply delivery status | | Extensibility | Packaged business skills, install-on-demand channel plugins, MCP servers, admin connector flows, and SecretRef-backed HTTP tools |

Product Strengths

• Validated business skills: production skills use deterministic helpers, fixtures, eval scenarios, and targeted tests so small models can perform useful work through structured actions.
• Guided onboarding: hatching collects useful work context, keeps setup links handy, writes first-job suggestions, and can route the first-run conversation through a stronger onboarding model before returning to the normal default. Its lifecycle is searchable in audit as onboarding events.
• Multi-agent operations: agents can coordinate across local workspaces, hosted HybridAI proxies, and trusted peer HybridClaw instances with A2A pairing, pinned X25519 keys, encrypted envelopes, explicit addressing, inbound messages, reply-back delivery status, admin-visible trust, and an A2A local mode that closes unrelated external surfaces.
• Prompt-level credential isolation: encrypted secrets and SecretRefs keep credential values out of model context while tools receive only the scoped credential material needed at execution time.
• One runtime for many surfaces: web, terminal, Discord, Slack, Teams, email, voice, webhooks, and local API clients all use the same gateway, memory, policy, and audit model.
• Generated apps from chat: describe an app, dashboard, document, game, or tool in web chat; HybridClaw saves the generated HTML artifact into the Apps gallery, refreshes live apps through read-only connector calls, and can share apps through expiring links or Microsoft Teams tabs.
• Secure by default: LLM output is treated as untrusted, risky actions route through approval policy, and every sensitive boundary is visible in audit.
• Model freedom: use HybridAI, the direct OpenAI Responses API, Codex OAuth, other hosted providers, local engines, or named OpenAI-compatible endpoints from the same model picker and config surface. Ordered routing tiers can start routine work cheaply and escalate retry-safe failures while explicit model pins remain authoritative.
• Operator visibility: /admin groups Activity, Agents, Automation, Connectivity, Models, Security, System, and Labs workflows. Page and setting search opens with Cmd/Ctrl+K, settings link to one canonical owner, and legacy admin URLs redirect to their corresponding tab.
• Business-ready extension model: packaged skills, plugins, MCP servers, and SecretRef-backed HTTP tools share the same approval and credential boundaries.
• Practical migration path: preview compatible imports from OpenClaw or Hermes, then package agents as portable .claw archives. HybridClaw is an independent implementation, not a fork — see the code provenance statement.

Common Commands

hybridclaw gateway status
hybridclaw tui --resume <sessionId>
hybridclaw config get <key>
hybridclaw token list
hybridclaw skill list
hybridclaw agent list
hybridclaw doctor
hybridclaw update --yes

Migration preview:

hybridclaw migrate openclaw --dry-run
hybridclaw migrate hermes --dry-run

HybridAI Platform

HybridClaw runs self-hosted. HybridAI is the optional platform layer around it:

• managed HybridClaw launch at hybridclaw.io
• enterprise shared RAG and cloud memory
• managed access to current models
• observability across multiple agents
• hosted email addresses for agents
• ready-to-run virtual coworkers

Architecture

User message
  -> Gateway (HTTP, web chat, TUI, Discord, Slack, email, etc.)
  -> ContainerInput JSON
  -> Host or Docker runtime
  -> Agent loop, tools, approvals, memory, MCP
  -> ContainerOutput JSON
  -> Gateway response, session store, audit trail

Core pieces:

• Gateway service: command handling, channel transports, REST APIs, scheduler, SQLite persistence, audit, A2A, admin surfaces.
• Container runtime: sandboxed tool execution, provider adapters, browser automation, media/search tooling, file-based IPC.
• TUI client: thin HTTP client for terminal-first operation.
• Console: web chat, Apps gallery, and admin UI served by the gateway.

Docs By Goal

| Goal | Start here | | --- | --- | | Install and launch | Quick Start, Installation | | Configure providers and models | Authentication, Model Selection | | Build generated apps | Apps Gallery | | Connect channels | Connect Your First Channel, Channels | | Use bundled skills | Bundled Skills, Skills Catalog | | Distill a coworker | Human Distillation | | Operate securely | Security, Trust Model, Approvals | | Inspect commands | Commands, Diagnostics | | Extend HybridClaw | Extensibility, Plugins, MCP | | Build desktop releases | Desktop Release Builds | | Contribute | CONTRIBUTING.md, docs/content/README.md |

Latest release: v0.28.5. Release notes: CHANGELOG.md

Development

npm install
npm run setup
npm run build
npm run typecheck
npm run test:unit

Useful dev commands:

npm run dev          # gateway in hot-reload mode
npm run tui          # terminal client
npm run check        # Biome check
npm run format       # Biome write

For docs-only changes, verify links, commands, and examples. For code changes, run npm run typecheck, npm run lint, and targeted tests for the touched area.

Community

• Discord: discord.gg/jsVW4vJw27
• Issues: github.com/HybridAIOne/hybridclaw/issues
• Discussions: github.com/HybridAIOne/hybridclaw/discussions
• Support guide: SUPPORT.md
• Community standards: CODE_OF_CONDUCT.md