QwryAI
/install qwryai
QwryAI Agent Access
Use this skill when the user wants OpenClaw to inspect a QwryAI workspace through QwryAI's read-only developer and agent integration surface.
What This Connects
QwryAI exposes active-paid-subscription-gated, tenant-isolated access through:
- MCP:
https://api.qwryai.com/mcp - Public REST API:
https://api.qwryai.com/public/v1 - Claude OAuth MCP:
https://api.qwryai.com/mcp-oauth - CLI:
qwryai-cli
This skill is read-only. It can help agents list chatbots, inspect conversations, read messages, search knowledge, and view analytics. It must not send replies, modify chatbots, change billing, manage users, or perform admin actions. ClawHub does not bill for this skill; QwryAI enforces access through the customer's current paid subscription plan.
Required Setup
- In QwryAI, open Settings > Security > Developer & Agent Access.
- Confirm the workspace has an active paid QwryAI subscription with developer access enabled.
- Generate a read-only API key.
- Set:
export QWRYAI_API_KEY=\x3Cyour-qwryai-api-key>
export QWRYAI_MCP_URL=https://api.qwryai.com/mcp
export QWRYAI_API_URL=https://api.qwryai.com/public/v1
OpenClaw MCP Definition
Add QwryAI as a remote HTTPS MCP server in OpenClaw. QwryAI does not require a local stdio process; clients should call the hosted MCP endpoint with a bearer API-key header.
{
"url": "https://api.qwryai.com/mcp",
"headers": {
"Authorization": "Bearer ${QWRYAI_API_KEY}"
}
}
If the OpenClaw runtime does not interpolate environment variables in MCP headers, paste the generated API key directly into the local OpenClaw configuration and keep that config out of source control.
Claude Uses OAuth
Claude should use the OAuth MCP endpoint:
https://api.qwryai.com/mcp-oauth
Do not paste a QwryAI API key into Claude. Claude should open QwryAI sign-in and approval, then receive an OAuth token scoped to the approved read-only access.
Cursor And Custom MCP Definition
For Cursor-style MCP configuration, add QwryAI as a remote HTTPS MCP server:
{
"mcpServers": {
"qwryai": {
"url": "https://api.qwryai.com/mcp",
"headers": {
"Authorization": "Bearer ${QWRYAI_API_KEY}"
}
}
}
}
Compatible custom MCP clients can use the same URL and bearer-token header.
Available Read-Only Tools
get_account_contextlist_chatbotsget_chatbotlist_conversationsget_conversationget_conversation_messagessearch_knowledgeget_analytics_overview
Security Notes
- This skill requires an active paid QwryAI subscription with Developer & Agent Access enabled. ClawHub does not bill for this skill; QwryAI subscription access is enforced by QwryAI.
- Do not add ClawHub paid-skill, pricing, or revenue-share metadata. Billing stays inside the customer's existing QwryAI plan.
- Registry metadata must declare
QWRYAI_API_KEYas required and sensitive. Optional endpoint overrides areQWRYAI_MCP_URLandQWRYAI_API_URL. - Network access is limited to
https://api.qwryai.com. - Treat
QWRYAI_API_KEYas a secret. - Never paste QwryAI API keys into prompts, tickets, public logs, or shared screenshots.
- Access follows the tenant's current active subscription plan. If the workspace is downgraded, canceled, inactive, suspended, or using revoked or expired API/OAuth credentials, API, MCP, OAuth, and CLI access should stop.
- The V1 integration surface is intentionally read-only.
- Do not ask the agent to send replies, resolve conversations, create/update/delete resources, manage users, or manage billing.
Useful Prompts
- "List my QwryAI chatbots and summarize which ones have recent conversations."
- "Search QwryAI knowledge for the refund policy in chatbot
\x3Cid>." - "Show analytics overview for the last 30 days."
- "Read the latest messages in conversation
\x3Cid>and summarize customer intent."
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install qwryai - 安装完成后,直接呼叫该 Skill 的名称或使用
/qwryai触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
QwryAI 是什么?
Connect OpenClaw to a paid QwryAI workspace through the read-only QwryAI MCP and public API integration surface. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 62 次。
如何安装 QwryAI?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install qwryai」即可一键安装,无需额外配置。
QwryAI 是免费的吗?
是的,QwryAI 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
QwryAI 支持哪些平台?
QwryAI 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 QwryAI?
由 Tamjeed Hur(@tamjeedhur)开发并维护,当前版本 v0.1.0。