QVeris
Run a task

QVeris Open Ecosystem — Frequently Asked Questions

The QVeris ecosystem includes Hosted MCP (recommended for remote-capable MCP clients), the CLI (v0.11.0, recommended for shell-capable agents), Local MCP Server (v0.14.0, stdio fallback), TypeScript SDK (v0.8.0), Python SDK (v0.7.0), and community-contributed skills distributed through GitHub, npm, and ClawHub. All developer-facing tooling is published openly for inspection and reuse.

What is the QVeris open ecosystem?
The QVeris ecosystem is a collection of open-source clients, SDKs, an MCP Server, CLI tools, and community-contributed skills that connect to the QVeris capability routing network. Source code is published on GitHub (github.com/QVerisAI), packages on npm (@qverisai), and skills on ClawHub.
What open-source tools does QVeris provide?
QVeris provides Hosted MCP as the preferred remote Streamable HTTP connection for compatible clients, plus the CLI (@qverisai/cli v0.11.0) for zero-schema-token agent tool calling and Local MCP Server (@qverisai/mcp v0.14.0) as the stdio fallback. Both MCP paths expose discover / inspect / probe / call plus usage_history and credits_ledger audit tools. It also publishes a TypeScript SDK (@qverisai/sdk v0.8.0), a Python SDK (v0.7.0), a REST API, and the OpenClaw capability runtime. All packages are available on GitHub and npm under the QVerisAI organization.
What is ClawHub and how does it relate to QVeris?
ClawHub is a skill distribution platform for AI agent capabilities. QVeris publishes reusable skills to ClawHub, and agents can install them to extend their capability set. ClawHub works alongside GitHub and npm as part of the QVeris open ecosystem.
How can I contribute to the QVeris ecosystem?
Developers can contribute by publishing skills to ClawHub, submitting pull requests to QVeris GitHub repositories, creating npm packages that extend QVeris functionality, or building new capability providers that integrate with the routing network.
What is OpenClaw and how does it relate to QVeris?
OpenClaw is an open-source capability runtime that QVeris contributes to upstream. OpenClaw handles execution of individual capabilities, while QVeris provides the managed routing network layer — discovery, quality signals, provider selection, and billing — on top of it.
Ecosystem

Built together,
open by default.

QVeris is designed to work with the tools agents and developers already use. Open-source CLI, typed SDKs, MCP server, and a growing community of skill builders.

QVeris Products

Built and maintained by QVeris AI

Two first-party open-source projects for different use cases — a token-efficient CLI for agents and a multi-agent workflow engine. Both have native QVeris integration out of the box.

By QVeris AI

QVeris CLI

Zero-token tool calling for agents and developers

Discover, inspect, and call 10,000+ real-world, verified capabilities from the terminal. Runs as a subprocess — zero schema tokens, schemas never enter the LLM context. Interactive REPL for human exploration, --json structured output for agent automation, qveris doctor self-check, qveris usage / qveris ledger billing audit. Zero external dependencies.

  • ▸Zero schema tokens vs MCP — runs outside the LLM context, tool registries stay out of the prompt
  • ▸Interactive REPL with colored output, spinners, and codegen (curl / Python / JS)
  • ▸Agent-native: --json output, probe validation and quoting, --dry-run preview, structured exit codes
  • ▸One-line site-aware install + interactive login: curl -fsSL https://qveris.ai/cli/install | bash; then qveris login
  • ▸Session index shortcuts (30-min TTL) + qveris usage / qveris ledger for charge_outcome audit
Best for: Agents with shell access (Claude Code, OpenClaw, Hermes Agent, terminals) that need zero-schema-token tool calling, and developers who want terminal-first API discovery with built-in billing audit.
View on GitHub →
By QVeris AI

QVerisFlow

Multi-agent workflow engine powered by LangGraph + QVeris

A LangGraph-based multi-agent framework for building, visualizing, and executing complex AI workflows. Create workflows through natural language conversation or a visual canvas editor. A Meta Agent automatically analyzes tasks and generates optimal agent definitions with QVeris as the unified data and tool layer.

  • ▸Conversational workflow building — describe your need in natural language, get a working workflow
  • ▸Visual canvas editor (React + ReactFlow) with drag-and-drop, real-time state, and live execution monitoring
  • ▸Meta Agent auto-generates agent definitions and orchestration from task descriptions
  • ▸Multi-model support: Qwen, DeepSeek, GPT, GLM, Kimi, and more
  • ▸Three-layer verification: rule validation + LLM evaluation + composite checks
Best for: Developers building complex multi-agent pipelines — financial analysis, research automation, data processing — where multiple specialized agents collaborate through structured workflows.
View on GitHub →
Agent Clients

Agent Clients

Plugin + Skill

OpenClaw

The open-source AI agent client. Integrate QVeris via the official Plugin (recommended) or Skill — install with one command, then configure your API key.

GitHub →Full setup guide →
Recommended

Plugin — via npm

Runtime-level tool registration. Tools stay available regardless of context length — no prompt competition, no forgetting.

1. Install
openclaw plugins install @qverisai/qveris
Copy prompt (log in to include your key) ↓
Install QVeris plugin: openclaw plugins install @qverisai/qveris, then configure plugins.allow: ["qveris"], plugins.entries.qveris: { enabled: true, config: { apiKey: "YOUR_KEY" } }, and tools.alsoAllow: ["qveris"] in openclaw.json
View on npm →
Lightweight alternative

Skill — via ClawHub

Instruction-only, zero dependencies. Good for quick trials or restricted environments where npm is unavailable.

1. Install
openclaw skills install qveris-official
Copy prompt (log in to include your key) ↓
Install QVeris skill: openclaw skills install qveris-official, then add "qveris-official": { "enabled": true, "apiKey": "YOUR_KEY" } to skills.entries in ~/.openclaw/openclaw.json
View on ClawHub →
MCP compatible

Cursor

AI-powered code editor. Add QVeris via MCP configuration to give Cursor agents access to 10,000+ real-world, verified capabilities.

Learn more →
CLI / MCP

Claude Code

Anthropic's agentic coding tool. Use QVeris CLI (recommended) or install the MCP server to extend Claude with real-world data and tools.

Learn more →
CLI / MCP

OpenCode

Open-source terminal coding agent. Add QVeris MCP to opencode.json or use the CLI for low-token tool access.

Learn more →
SDKs & Packages

SDKs & Packages

PyPI

Python SDK

v0.7.0 — pip install qveris. Full typed API surface, async support, retry logic, and streaming. The standard way to integrate QVeris in Python agents.

Learn more →
npm

npm Packages

@qverisai/sdk v0.8.0 and other packages for JavaScript/TypeScript environments. Works in Node.js, Deno, and edge runtimes.

Learn more →
MCP

@qverisai/mcp

v0.14.0 — MCP server exposing discover, inspect, probe, call plus usage_history and credits_ledger audit tools. For Cursor, Claude Desktop, and all MCP-compatible agents. Configure QVERIS_BASE_URL with the endpoint shown by this site.

Learn more →
Community & Resources

Community & Resources

Community

ClawHub Skills

Community-curated skill packs for OpenClaw and Claude. Browse, install, and share agent skills that use QVeris capabilities.

Learn more →
Open source

GitHub Organization

All QVeris open-source repositories: CLI, MCP server, Python SDK, example agents, skills, and contributing guides.

Learn more →
Documentation

REST API Reference

Full Discover, Inspect, and Call endpoint documentation with request/response examples, authentication guide, and error codes.

Learn more →

Join the ecosystem

Get an API key, build with the CLI or SDK, and contribute back. Everything you build on QVeris can be shared as a skill on ClawHub.

Get started — freeGitHub →