Agent 配套工具

Helicone｜LLM 可观测性与 AI 网关平台

Helicone

面向 AI 工程师的开源 LLM 可观测性与 AI 网关平台，可记录并分析模型请求的调用链路、成本与延迟，并通过统一接口访问多家模型服务。

工程开发 API集成审计与回放多模型接入
获取与使用

需要部署

按官方资料自行安装或部署

获取与使用 查看 GitHub ↗ 访问官方网站 ↗
使用难度待核验
国内可用性待核验

2026-11-09 更新 · 本站尚未实测

项目介绍

Agent 配套工具 · 2026-11-09 更新 · 本站尚未实测

Helicone 是一个面向 AI 工程师的 AI Gateway 与 LLM Observability 平台。按仓库 README 的定位，它一方面做可观测：记录并检查 agent、聊天机器人、文档处理流水线等的 traces 与 sessions；另一方面做网关：通过 OpenAI 兼容接口访问 100+ AI 模型，并提供智能路由与自动 fallback。

需要明确的是，它是一个平台型项目，而不是可下载的单个 Agent 或工作流文件。仓库主体为 TypeScript，README 列出的自托管架构包含 Web 前端、Worker（Cloudflare Workers 代理日志）、Jawn 日志收集服务、Supabase 数据库与鉴权、ClickHouse 分析库、Minio 对象存储等服务。

核心能力

点击能力名称查看输入、输出和原文依据。

  • AI Gateway：README 称可通过一个 API Key、以 OpenAI API 形式访问 100+ 模型，支持智能路由与自动 fallback。
  • 快速接入：README 称一行代码即可记录来自 OpenAI、Anthropic、LangChain、Gemini、Vercel AI SDK 等的请求。
  • 可观测与调试：检查并调试 agent、聊天机器人、文档处理流水线的 traces 与 sessions。
  • 指标分析：跟踪成本、延迟、质量等指标，并可一行导出到 PostHog 做自定义看板。
  • Playground 与提示词管理：在界面中测试迭代 prompt、session 与 trace；对 prompt 做版本管理，并称可通过 AI Gateway 在不改代码的情况下部署 prompt。
  • 微调：README 称可与 OpenPipe、Autonomi 等微调合作方配合使用。

README 还提到 LLM Cost API（称有 300+ 模型与供应商的定价数据库），以及可通过 API 或 MCP server 管理和导出数据。

实际应用场景

适用场景

  • 需要在生产环境中观察 LLM 应用调用链路、排查异常的团队；
  • 希望统计模型调用成本与延迟、做多供应商对比的团队；
  • 想在多个模型供应商之间做统一接入与故障切换的项目；
  • 需要对 prompt 做版本管理与在线迭代的团队。

如果你的需求只是单次对话或简单脚本调用，本项目的平台属性可能偏重，需评估自托管组件数量。

安装与接入

使用托管服务（README 快速开始）：

  1. 在 Helicone 注册获取 API Key，并在其 credits 页面添加额度；
  2. 将代码中的 baseURL 改为 https://ai-gateway.helicone.ai，并填入 API Key；
  3. 在 Helicone 控制台查看日志。

自托管（README 给出 Docker 方式），原文命令为：

# Clone the repository
git clone https://github.com/Helicone/helicone.git
cd docker
cp .env.example .env

# Start the services
./helicone-compose.sh helicone up

README 还说明有面向企业负载的 Helm chart，需通过 enterprise@helicone.ai 联系获取；手动部署不被推荐，应优先使用 Docker 或 Helm。具体环境变量含义、端口、依赖版本等请以官方文档为准，本次资料未展开。

使用条件与限制

是否需要部署
需自行准备运行环境，具体要求见正文
是否需要注册
账号和认证要求见正文；未明确说明的项目仍待核验
是否需要 API Key
待核验
费用与额度
费用边界见正文，项目许可不代表模型或托管服务免费
所需权限
待核验
许可证与服务条款
待核验

权限费用与限制

  • README 称有每月免费额度（10k requests/month，无需信用卡），并在注册后需添加 credits；具体价格、额度上限与商业条款待向官方核验，本文不报具体价格。
  • 托管服务与自托管是两条不同路径：自托管需要自行准备数据库、对象存储等服务，运维成本由使用者承担。
  • 使用 AI Gateway 调用第三方模型时，模型调用费用、速率限制、数据处理条款取决于对应模型供应商，需另行核验。
  • README 称其企业版满足 SOC 2 与 GDPR 合规，该声明来自项目方，实际适用范围待核验。
  • 数据留存、日志内容是否包含敏感信息，需结合自身合规要求评估。

国内使用情况

网友核验 用户经验分享

与站长实测独立，不代表官方认证或本站已实测。

我要核验
0 位参与者0 份有效分项记录

最近核验：暂无有效核验。参与人数按当前项目有效身份去重；不同核验项分别统计，不合并成用户总结果。仅统计审核通过且未过期的记录。

分项统计、有效期与样本量
核验项目有效人数成功 / 部分 / 失败 / 无法判断成功率最近核验 / 有效期
官方网站访问00 / 0 / 0 / 0样本不足暂无 / 7 天
GitHub 仓库访问00 / 0 / 0 / 0样本不足暂无 / 7 天
安装与运行00 / 0 / 0 / 0样本不足暂无 / 30 天
官方服务国内连接 / MCP 连接00 / 0 / 0 / 0样本不足暂无 / 30 天
国内网络情况00 / 0 / 0 / 0样本不足暂无 / 7 天
国内模型兼容性00 / 0 / 0 / 0样本不足暂无 / 30 天
自定义 API00 / 0 / 0 / 0样本不足暂无 / 30 天
海外服务依赖
00 / 0 / 0 / 0样本不足暂无 / 30 天
海外账号要求
00 / 0 / 0 / 0样本不足暂无 / 30 天
中文界面
00 / 0 / 0 / 0样本不足暂无 / 30 天
客户端兼容性00 / 0 / 0 / 0样本不足暂无 / 30 天
账号及费用信息
00 / 0 / 0 / 0样本不足暂无 / 30 天

成功率 = 成功 ÷（成功 + 部分成功 + 失败）；无法判断不计入分母。至少 5 份可判断结果才显示成功率。账号、费用、海外依赖与语言等事实项，成功表示核查结论与资料吻合，具体观察另行统计；不表示免费或没有海外依赖。不同网络、客户端和版本的结果不能直接等同。

最近有效核验

暂无审核通过且未过期的网友核验。欢迎提交真实使用记录。

查看全部有效核验 →

国内实测状态：国内可用性待核验。协议兼容和官方资料说明不等于国内网络可用。

官方网站访问
尚未实测 待核验
GitHub 仓库访问
尚未实测 待核验
官方服务国内连接
尚未实测 待核验
国内网络情况
尚未实测 待核验
国内模型兼容性
尚未实测 待核验
自定义 API
尚未实测 待核验
海外服务依赖
尚未实测 待核验
海外账号要求
尚未实测 待核验
中文界面
尚未实测 待核验
国内最后核验时间
尚未实测

许可与国内使用

仓库代码采用 Apache License 2.0（见仓库 LICENSE 文件）。需要注意：公开仓库不等于全部服务免费，托管平台服务、商业模块、微调合作方与第三方模型服务各有自己的条款，需分别核验。

README 列出的模型供应商多为境外服务（OpenAI、Anthropic、Gemini、AWS Bedrock、Groq 等），托管入口为境外域名；国内网络可达性、中文模型支持情况本次资料未说明，待核验。若需在境内使用，建议优先评估自托管路径与所用模型服务的合规与网络条件。

项目资料与来源

官方网站 ↗

GitHub 仓库 ↗

官方文档 ↗

字段依据与待核验事项

access：来源确认 · git clone https://github.com/Helicone/helicone.git cd docker cp .env.example .env

project_type：来源确认 · Helicone is an AI Gateway & LLM Observability Platform for AI Engineers

原始 README（折叠技术参考）
<div align="center">

| 🔍 Observability | 🕸️ Agent Tracing | 🚂 LLM Routing |
| :--------------: | :--------------: | :------------------: |
|  💰 Cost & Latency Tracking  |   📚 Datasets & Fine-tuning    |    🎛️ Automatic Fallbacks   |

</div>

<p align="center" style="margin: 0; padding: 0;">
  <img alt="helicone logo" src="https://marketing-assets-helicone.s3.us-west-2.amazonaws.com/Twitter_Cover_A1.png" style="display: block; margin: 0; padding: 0;">
</p>
</br>

<p align="center">
  <a href='https://github.com/helicone/helicone/graphs/contributors'><img src='https://img.shields.io/github/contributors/helicone/helicone?style=flat-square' alt='Contributors' /></a>
  <a href='https://github.com/helicone/helicone/stargazers'><img alt="GitHub stars" src="https://img.shields.io/github/stars/helicone/helicone?style=flat-square"/></a>
  <a href='https://github.com/helicone/helicone/pulse'><img alt="GitHub commit activity" src="https://img.shields.io/github/commit-activity/m/helicone/helicone?style=flat-square"/></a>
  <a href='https://github.com/helicone/helicone/issues?q=is%3Aissue+is%3Aclosed'><img alt="GitHub closed issues" src="https://img.shields.io/github/issues-closed/helicone/helicone?style=flat-square"/></a>
  <a href='https://www.ycombinator.com/companies/helicone'><img alt="Y Combinator" src="https://img.shields.io/badge/Y%20Combinator-Helicone-orange?style=flat-square"/></a>
</p>
<p align="center">
  <a href="https://docs.helicone.ai/">Docs</a> • <a href="https://www.helicone.ai/changelog">Changelog</a> • <a href="https://github.com/helicone/helicone/issues">Bug reports</a> • <a href="https://helicone.ai/demo">See Helicone in Action! (Free)</a>
</p>

## Helicone is an AI Gateway & LLM Observability Platform for AI Engineers

- 🌐 **AI Gateway**: Access 100+ AI models with 1 API key through the OpenAI API with intelligent routing and automatic fallbacks. [Get started in 2 minutes.](https://docs.helicone.ai/gateway/overview)
- 🔌 **Quick integration**: One-line of code to log all your requests from [OpenAI](https://www.helicone.ai/models?providers=openai), [Anthropic](https://www.helicone.ai/models?providers=anthropic), [LangChain](https://docs.helicone.ai/gateway/integrations/langchain), [Gemini](https://www.helicone.ai/models?providers=gemini%2Cgoogle-ai-studio), [Vercel AI SDK](https://docs.helicone.ai/gateway/integrations/vercel-ai-sdk), and [more](https://docs.helicone.ai/gateway/overview).
- 📊 **Observe**: Inspect and debug traces & [sessions](https://docs.helicone.ai/features/sessions) for agents, chatbots, document processing pipelines, and more
- 📈 **Analyze**: Track metrics like [cost](https://docs.helicone.ai/faq/how-we-calculate-cost#developer), latency, quality, and more. Export to [PostHog](https://docs.helicone.ai/getting-started/integration-method/posthog) in one-line for custom dashboards
- 🎮 **Playground**: Rapidly test and iterate on prompts, sessions and traces in our UI.
- 🧠 **Prompt Management**: [Version prompts](https://docs.helicone.ai/features/prompts) using production data. Deploy prompts through the AI Gateway without code changes. Your prompts remain under your control, always accessible.
- 🎛️ **Fine-tune**: Fine-tune with one of our fine-tuning partners: [OpenPipe](https://openpipe.ai/) or [Autonomi](https://www.autonomi.ai/) (more coming soon)
- 🛡️ **Enterprise Ready**: SOC 2 and GDPR compliant

> 🎁 Generous monthly [free tier](https://www.helicone.ai/pricing) (10k requests/month) - No credit card required!
>
<img src="https://github.com/user-attachments/assets/e16332e9-d642-427e-b3ce-1a74a17f7b2c" alt="Open Sourced LLM Observability & AI Gateway Platform" width="600">

## Quick Start ⚡️

1. Get your API key by signing up [here](https://helicone.ai/signup) and add credits at [helicone.ai/credits](https://us.helicone.ai/credits)

2. Update the `baseURL` in your code and add your API key.

   ```typescript
   import OpenAI from "openai";

   const client = new OpenAI({
     baseURL: "https://ai-gateway.helicone.ai",
     apiKey: process.env.HELICONE_API_KEY,
   });

   const response = await client.chat.completions.create({
     model: "gpt-4o-mini",  // claude-sonnet-4, gemini-2.0-flash or any model from https://www.helicone.ai/models
     messages: [{ role: "user", content: "Hello!" }]
   });
   ```

3. 🎉 You're all set! View your logs at [Helicone](https://us.helicone.ai/dashboard) and access 100+ models through one API.

### Self-Hosting Open Source LLM Observability

#### Docker

Helicone is simple to self-host and update. To get started locally, just use our [docker-compose](https://docs.helicone.ai/getting-started/self-deploy-docker) file.

```bash
# Clone the repository
git clone https://github.com/Helicone/helicone.git
cd docker
cp .env.example .env

# Start the services
./helicone-compose.sh helicone up
```

#### Helm

For Enterprise workloads, we also have a production-ready Helm chart available. To access, contact us at enterprise@helicone.ai.

#### Manual (Not Recommended)

Manual deployment is not recommended. Please use Docker or Helm. If you must, follow the instructions [here](https://docs.helicone.ai/getting-started/self-deploy).

#### Architecture

Helicone is comprised of five services:

- **Web**: Frontend Platform (NextJS)
- **Worker**: Proxy Logging (Cloudflare Workers)
- **Jawn**: Dedicated Server for serving collecting logs (Express + Tsoa)
- **Supabase**: Application Database and Auth
- **ClickHouse**: Analytics Database
- **Minio**: Object Storage for logs.

## Integrations 🔌

### Inference Providers

| Integration                                                                            | Supports                                                                                                                                     | Description                                           |
| -------------------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------- |
| AI Gateway                                | [JS/TS, Python, cURL](https://docs.helicone.ai/gateway/overview)                                                                                                                          | Unified API for 100+ providers with intelligent routing, automatic fallbacks, and unified observability
| Async Logging (OpenLLMetry)                                                            | [JS/TS](https://docs.helicone.ai/getting-started/integration-method/openllmetry), [Python](https://www.npmjs.com/package/@helicone/helicone) | Asynchronous logging for multiple LLM platforms       |
| OpenAI                                                                                 | [JS/TS, Python](https://www.helicone.ai/models?providers=openai)              | Inference provider                                                     |
| Azure OpenAI                                                                           | [JS/TS, Python](https://www.helicone.ai/models?providers=azure)                | Inference provider                                                     |
| Anthropic                                                                              | [JS/TS, Python](https://www.helicone.ai/models?search=anthropic)        | Inference provider                                                     |
| Ollama                                                                                 | [JS/TS](https://docs.helicone.ai/integrations/ollama/javascript)                                                                             | Run and use large language models locally             |
| AWS Bedrock                                                                            | [JS/TS](https://www.helicone.ai/models?providers=azure%2Cbedrock)                                                                            | Inference provider                                                     |
| Gemini API                                                                             | [JS/TS](https://www.helicone.ai/models?providers=google-ai-studio)                                                                         | Inference provider                                                     |
| Gemini Vertex AI                                                                       | [JS/TS](https://www.helicone.ai/models?providers=vertex)                                                                      | Gemini models on Google Cloud's Vertex AI             |
| Vercel AI                                                                              | [JS/TS](https://docs.helicone.ai/gateway/integrations/vercel-ai-sdk)                                                                           | AI SDK for building AI-powered applications           |
| Anyscale | [JS/TS, Python](https://www.helicone.ai/models?providers=anyscale)                                                                                                                                | Inference provider                                                     |
| TogetherAI | [JS/TS, Python](https://www.helicone.ai/models?providers=together)     | Inference provider                                                                                                                                | -                                                     |
| Hyperbolic | [JS/TS, Python](https://www.helicone.ai/models?providers=hyperbolic)   | Inference provider                                                                                                                                | High-performance AI inference platform                |
| Groq                                                                                   | [JS/TS, Python](https://www.helicone.ai/models?providers=groq)                  | High-performance models                               |
| DeepInfra     | [JS/TS, Python](https://www.helicone.ai/models?providers=deepinfra)                                                                                                                                | Serverless AI inference for various models            |       |
| Fireworks AI  | [JS/TS, Python](https://www.helicone.ai/models?providers=fireworks)                                                                                                                                | Fast inference API for open-source LLMs               |

### Frameworks

| Framework                                                             | Supports                                                            | Description                                                                             |
| --------------------------------------------------------------------- | ------------------------------------------------------------------- | --------------------------------------------------------------------------------------- |
| LangChain   | [JS/TS, Python](https://www.helicone.ai/models?providers=langchain)                                                       | Use AI Gateway with LangChain for unified provider access                               |
| LlamaIndex | [Python](https://www.helicone.ai/models?providers=llamaindex)                                                              | Framework for building LLM-powered data applications                                    |
| LangGraph   | [Python](https://www.helicone.ai/models?providers=langgraph)                                                              | Build stateful, multi-actor applications with LLMs                                       |
| Vercel AI SDK | [JS/TS](https://www.helicone.ai/models?providers=vercel-ai-sdk)                                                               | AI SDK for building AI-powered applications                                              |
| Semantic Kernel | [C#, Python](https://www.helicone.ai/models?providers=semantic-kernel)                                                          | Microsoft's AI orchestration framework                                                  |
| CrewAI         | [Python](https://docs.helicone.ai/integrations/openai/crewai)                                                                   | Framework for orchestrating role-playing AI agents                                      |                                                           |
| ModelFusion                            | [JS/TS](https://github.com/vercel/modelfusion/blob/main/docs/integration/observability/helicone.md) | Abstraction layer for integrating AI models into JavaScript and TypeScript applications |
| PostHog | [JS/TS, Python, cURL](https://docs.helicone.ai/getting-started/integration-method/posthog) | Product analytics platform. Build custom dashboards.    |
| RAGAS                     | [Python](https://docs.helicone.ai/other-integrations/ragas) | Evaluation framework for retrieval-augmented generation |
| Open WebUI           | [JS/TS](https://docs.helicone.ai/other-integrations/open-webui) | Web interface for interacting with local LLMs           |
| MetaGPT                | [YAML](https://docs.helicone.ai/other-integrations/meta-gpt) | Multi-agent framework                                   |
| Open Devin           | [Docker](https://docs.helicone.ai/other-integrations/open-devin) | AI software engineer                                    |
| Mem0 EmbedChain      | [Python](https://docs.helicone.ai/other-integrations/embedchain) | Framework for building RAG applications                 |
| Dify                      | [No code required](https://docs.helicone.ai/other-integrations/dify) | LLMOps platform for AI-native application development   |

> This list may be out of date. Don't see your provider or framework? Check out the latest integrations in our [docs](https://docs.helicone.ai/gateway/integrations/overview). If not found there, request a new integration by contacting help@helicone.ai.

## Contributing

We ❤️ our contributors! We warmly welcome contributions for documentation, integrations, costs, and feature requests.

If you have an idea for how Helicone can be better, create a [GitHub issue](https://github.com/Helicone/helicone/issues).

## License

Helicone is licensed under the [Apache v2.0 License](LICENSE).

## Additional Resources

- **LLM Cost API**: We have the largest open-source API pricing database with 300+ models and providers such as OpenAI, Anthropic and more. [Start querying here.](https://www.helicone.ai/llm-cost)

- **Data Management**: Manage and export your Helicone data with our [API](https://docs.helicone.ai/rest/user/post-v1userquery) or access it with our [MCP server](https://docs.helicone.ai/integrations/tools/mcp).

  - Guides: [ETL](https://docs.helicone.ai/use-cases/etl), [Request Exporting](https://docs.helicone.ai/use-cases/getting-user-requests)

- **Data Ownership**: Learn about [Data Ownership and Autonomy](https://docs.helicone.ai/use-cases/data-autonomy)

For more information, visit our [documentation](https://docs.helicone.ai/).

# Contributors

<a href="https://github.com/Helicone/helicone/graphs/contributors">
  <img src="https://contrib.rocks/image?repo=Helicone/helicone" />
</a>
历史前置条件说明（补充参考）

注册 Helicone 获取 API Key（托管网关需要）；自托管需 Docker 环境，并自行准备 Supabase、ClickHouse、Minio 等依赖；Helm 方式需联系官方获取；调用第三方模型产生的费用与额度需向对应供应商核验；网络可达性、国内模型支持待核验。

项目来源与清单版权许可

https://www.helicone.ai/

https://github.com/Helicone/helicone

整理来源：https://github.com/slavakurilyak/awesome-ai-agents

清单版权与许可：

MIT License

Copyright (c) 2024 Slava Kurilyak

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.