Okareo

com.okareo由 okareo.com 发布官方发布者v0.0.57更新于 Oct 27, 2026

Simulation, evaluation and monitoring for voice agents.

已验证Streamable HTTP可网页运行安全与监控数据与分析AI 与机器学习

概览

AI 生成的概览

Okareo 将语音智能体的模拟、评估与监控平台以 MCP 工具形式提供,让助手可以创建场景、运行测试与模拟并查看结果。

可以做什么
  • 创建、版本化、查看和删除由输入/结果行组成的场景
  • 注册和管理生成模型与模拟目标,包括语音目标
  • 运行质量测试和多轮模拟,读取运行摘要、逐项检查结果和对话记录
  • 创建、生成、校准和删除质量检查
  • 管理语音服务商集成,并提交已完成的语音对话用于监控
  • 管理项目与租户,查询分析数据,保存或调整仪表板顺序
适用场景
适合在构建或维护语音、LLM 智能体时,直接在编辑器中做评估:编写测试场景、用人物设定运行模拟、用检查项给输出打分并跟踪结果变化。也适合在 Okareo 中监控线上语音流量的团队。
运行要求
远程端点 tools.okareo.com,使用 streamable HTTP,首次连接通过浏览器 OAuth 登录;较旧或无头客户端可改用 Okareo API 密钥作为 Authorization bearer 头(OKAREO_API_KEY)。需要 Okareo 账号。可选本地安装通过 uvx 或 pip,需 Python 3.10-3.12 并设置 OKAREO_API_KEY。
安装前请注意
需要 Okareo 账号;使用 bearer 回退方式时需提供可访问你项目与数据的 API 密钥,建议用环境变量而非明文写入配置。许多工具会写入或删除:场景、检查项、模型、目标、人物设定、仪表板和语音集成都可能被永久删除,运行测试与模拟会消耗 Okareo 用量。轮换语音集成密钥会改变语音服务商使用的凭据。

安装

在 SourceWeft 中

  1. 打开 控制台中的 Okareo,将其添加到工作区。
  2. 连接 所需的 凭据。它们会被私密存储,绝不会在公开页面上展示。
  3. 为需要使用其工具的对话启用该服务。

Web executable,通过 Streamable HTTP。 远程服务在工作区中配置后即可从网页运行时运行。

交给智能体安装

把这段提示词粘贴到 Claude Code、Codex、Cursor 等智能体,或 SourceWeft 对话中。智能体会读取该服务器的安装指南,先向你展示发布者、端点和登录方式,经你同意后再把它添加到你的客户端。密钥不会写进任何命令。

阅读 https://sourceweft.com/mcp/com.okareo%2Fokareo/install.md,并按其中说明把这个 MCP 服务器接入我的客户端。修改任何配置之前,先告诉我它的发布者、端点和登录方式,等我确认。不要把 API 密钥写进命令或我会提交的文件;问我密钥应从哪里获取。

查看智能体所依据的安装指南

其他 MCP 客户端

把它添加到你客户端的 mcpServers 配置中。 客户端会在首次连接时完成 OAuth 登录。

{
  "mcpServers": {
    "okareo": {
      "type": "http",
      "url": "https://tools.okareo.com/mcp"
    }
  }
}

README

Okareo MCP Server

The Okareo MCP server exposes Okareo's evaluation capabilities as MCP tools, allowing AI coding assistants to create and manage scenarios, register models, run evaluations, and execute multi-turn simulations directly from your editor.

For detailed documentation, see the Okareo MCP docs.

Prerequisites

  • An Okareo account at app.okareo.com
  • A copilot that supports MCP servers (Claude Code, Cursor, VS Code, or GitHub Copilot), or Slackbot in Slack
  • Python 3.10–3.12 (only for the local install modes; not needed for remote)

Remote MCP (hosted) — recommended

The fastest way to get started. No local install, no Python, no API key in .mcp.json. Browser sign-in handles auth on first connect.

Recommended (OAuth — Claude Code, Claude Desktop, Cursor, VS Code 1.101+)

Add to your copilot's MCP config (typically .mcp.json):

json
{  "mcpServers": {    "okareo": {      "type": "http",      "url": "https://tools.okareo.com/mcp"    }  }}

Reload the copilot. It will open a browser to Okareo sign-in once; thereafter the copilot stores the token itself. Your .mcp.json contains no secrets.

Fallback (Bearer header — older clients or headless / CI)

For clients that haven't shipped the MCP OAuth flow yet, paste your API key as a bearer header. Prefer the env-var form over an inline literal:

json
{  "mcpServers": {    "okareo": {      "type": "http",      "url": "https://tools.okareo.com/mcp",      "headers": {        "Authorization": "Bearer ${env:OKAREO_API_KEY}"      }    }  }}

GitHub Copilot (in VS Code and the Copilot CLI), Slackbot and every other client are set up at docs.okareo.com/mcp/configuration.

Working across multiple Okareo organizations

If you have access to more than one Okareo account, simply Clear Authentication and then re-authenticate with the account you want to work in. You can always ask which account your are in and have acccess to.


Local install (alternative)

Run the MCP server on your own machine. Useful for offline / airgapped environments and for development.

Prefer the Remote MCP section above unless you have a specific reason to install locally (airgapped, custom build, development on this repo). The remote endpoint requires no install and stays current automatically.

For multi-org users, the remote endpoint also exposes tenant management tools (list_tenants, switch_tenant).

Step 1: Set Your API Key

bash
export OKAREO_API_KEY="your-api-key"

Add this to your ~/.zshrc or ~/.bash_profile for persistence.

Step 2: Configure Your Copilot

Claude Code

Add to .mcp.json:

json
{  "mcpServers": {    "okareo": {      "command": "uvx",      "args": ["okareo-mcp"],      "env": {        "OKAREO_API_KEY": "${OKAREO_API_KEY}"      }    }  }}

No pre-install needed — uvx handles it automatically.

Cursor

Add to .cursor/mcp.json:

json
{  "mcpServers": {    "okareo": {      "command": "uvx",      "args": ["okareo-mcp"]    }  }}

Cursor inherits OKAREO_API_KEY from your shell environment.

Alternative: pip install

If you don't have uv installed:

bash
pip install okareo-mcp

Then use "command": "okareo-mcp" instead of "command": "uvx" with "args": ["okareo-mcp"].


Configuration Reference

VariableDefaultDescription
OKAREO_API_KEY(required)Your Okareo API key
OKAREO_BASE_URLhttps://api.okareo.comOverride for on-prem Okareo backend
TRANSPORTstdioTransport: stdio (local) or sse (Docker)
PORT8000Port for SSE transport

Available Tools

Scenarios

ToolDescription
save_scenarioSave a named scenario from rows of input/result data (idempotent)
list_scenariosList all scenarios in the project with names, IDs, and row counts
get_scenarioRetrieve a scenario's metadata and all data rows by name or ID
create_scenario_versionCreate a new version of an existing scenario with updated data
preview_delete_scenarioPreview what will be deleted before removing a scenario
delete_scenarioPermanently delete a scenario and all related test data

Generation Models

ToolDescription
list_available_llmsBrowse available LLMs from the Okareo registry
register_generation_modelRegister a generation model for testing by selecting an LLM from the registry
list_generation_modelsList all registered generation models in the project
get_generation_modelRead detailed information about a registered generation model
update_generation_modelChange the LLM a registered generation model points to
delete_generation_modelRemove a registered generation model and all its related test data

Tests & Checks

ToolDescription
list_checksList available quality checks (built-in and custom) for evaluating model outputs
run_testRun a quality test that evaluates a model against a scenario using specified checks
list_test_runsFind past test runs, most recent first. detail_level="summary" (default) or "detailed" for aggregate metrics
get_test_run_resultsLoad one page of a run's results by ID or name. detail_level="summary" (default) / "detailed" (per-check outcomes + explanations) / "full" (transcripts). Defaults to 20 conversations. Carries a rerun block — the call and config to re-run this simulation
get_conversation_transcriptRetrieve the full conversation transcript for a single data point
reevaluate_test_runRe-score a completed test run against a (possibly different) set of checks
create_or_update_checkCreate or update a quality check by name — model-based, code-based, or audio (upsert)
generate_checkGenerate a check from a natural-language description, then save it
get_checkRetrieve a check's full configuration, including its prompt template or code
delete_checkPermanently delete a check by name
calibrate_checkDry-run a draft check against a finished test run — per-row verdicts and the arguments the check received, saving nothing

Simulations (Multi-Turn)

ToolDescription
create_or_update_targetCreate or update a Target — generation model, custom endpoint, or voice (OpenAI, Deepgram, Twilio)
get_targetRetrieve a Target's configuration by name (all types)
list_targetsFind simulation targets (voice and custom_endpoint). Bounded to 20 by default; name_contains filters
delete_targetRemove a simulation target and all its related test data
create_or_update_driverDefine a simulated user persona that will interact with your target
get_driverRetrieve a Driver's full configuration including the persona prompt
list_driversFind Driver personas. Bounded to 20 by default; name_contains filters
list_driver_voicesDiscover the voices, voice profiles, and languages available for voice drivers
run_simulationRun a multi-turn conversation evaluation. With based_on_run_id it inherits the original run's full configuration; any argument you pass overrides that one field
list_simulationsList past simulation runs with optional filters (target, scenario, limit)

Voice Monitoring

ToolDescription
ingest_conversationsSubmit completed voice conversations to Okareo for monitoring
connect_voice_integrationConnect a voice provider so its traffic flows into Okareo monitoring
list_voice_integrationsList the voice provider integrations in your project
get_voice_integrationRetrieve a voice provider integration by ID, including its status
update_voice_integrationUpdate a voice provider integration's metadata
rotate_voice_integration_secretRotate a voice provider integration's secrets
delete_voice_integrationDelete a voice provider integration by ID
get_voice_webhook_urlGet the inbound webhook endpoint for a voice provider

Projects

ToolDescription
list_projectsList every project you can work in — id, name, tags, archive state — with the active one marked
select_projectChoose the project to work in for this conversation
create_projectCreate a new project
update_projectRename a project, or replace its tags
archive_projectHide a project from the project picker. Reversible; nothing is deleted
unarchive_projectRestore an archived project to the picker
clone_projectCopy an existing project's scenarios into a new project (dry-run first)

Okareo has no project delete — in the app, the API, or here. Archiving is the removal, and it only hides the project from the picker: its scenarios, runs, and dashboards stay intact and every tool still works against it.

Analytics & Dashboards

ToolDescription
query_analyticsQuery Okareo's product analytics to understand evaluation trends
list_dashboardsList the analytics dashboards in your project
get_dashboardRetrieve a dashboard's full configuration by name
save_dashboardCreate or update an analytics dashboard by name (upsert)
reorder_dashboardsSet the display order of dashboards
delete_dashboardDelete a dashboard by name

Tenant Management (remote MCP only)

ToolDescription
list_tenantsList every Okareo organization you have access to in this MCP session
switch_tenantDoesn't change the active organization; returns how to reconnect and pick another one at sign-in

Documentation & Templates

ToolDescription
get_docsQuery the Okareo documentation system for conceptual or user-legible explanations
get_templatesRetrieve prompt templates for common Okareo patterns (works offline)
get_reps_baselineServe REPS agent-evaluation baseline material (scenario banks, drivers, checks, eval configs) from the latest tagged okareo-tools release

Troubleshooting

SymptomCauseFix
okareo-mcp: command not foundNot installed or not in PATHRun pip install -e . (dev) or use uvx okareo-mcp (user)
Server exits with API key errorOKAREO_API_KEY not setExport it: export OKAREO_API_KEY="..."
pip install fails on Python 3.13+Okareo SDK requires Python <3.13Use Python 3.10–3.12
Copilot can't connect (Docker)Wrong URLEnsure URL ends with /sse and port matches
Cursor doesn't pick up API keyCursor launched from Dock, not terminalLaunch Cursor from terminal: cursor .

Contributing

This repository is a curated public mirror; the canonical source is maintained by Okareo. We welcome issues and consider community pull requests — see CONTRIBUTING.md for how proposed changes are reviewed and ported.

License & Trademarks

The Okareo MCP server source code is licensed under the Apache License 2.0 (see also NOTICE).

"Okareo", the Okareo logo, and Okareo product names are trademarks of Okareo, Inc. and are not covered by the Apache 2.0 license. See TRADEMARK.md for permitted use.

来源:README.md,提交 f564678

工具

0
该服务器需要登录,因此 SourceWeft 无法在没有账号的情况下读取其工具,也未收录。已连接账号所能看到的内容可能因账号及其权限而异。

版本历史

7
  1. v0.0.57最新Oct 15, 2026
  2. vpublic-0.0.56Oct 14, 2026
  3. vpublic-0.0.55Oct 6, 2026
  4. vpublic-0.0.53Oct 5, 2026
  5. vpublic-0.0.52Sep 24, 2026
  6. vpublic-0.0.51Sep 23, 2026
  7. vpublic-0.0.50Sep 16, 2026

评分

已满足适用检查项的 85%

该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。

  • 近 30 天内可连通必需25 分
  • 已声明认证方式必需10 分
  • 已列出工具15 分

    请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。

  • 描述必需15 分
  • 图标10 分
  • 文档15 分
  • 许可证10 分