为什么值得关注

它结合了多模型支持、自主多步智能体和本地知识引擎，采用开放核心模式允许自托管和商业使用，仅限制直接竞争的主机服务。

适合谁使用

构建 AI 应用的开发者
管理大量文档的知识工作者和研究人员
希望本地运行模型或自托管的 AI 爱好者
需要私有、本地 AI 工作空间的组织

典型使用场景

使用多步智能体任务进行自动化研究和分析
构建并共享智能体生成的可复用工件
使用自定义模型对本地文档进行 RAG 问答
在统一界面中测试和对比多个 LLM

项目优势

支持多种模型，包括 Anthropic、OpenAI、Google Gemini 和本地 Ollama
开放核心许可允许自托管和商业衍生品
内置自主智能体，支持并行任务执行和人工干预
全面的 RAG 引擎，包含智能分块、向量搜索和集合管理

使用前须知

BUSL-1.1 许可证禁止将该软件作为竞争性托管服务提供
自托管需要 Docker 安装和多个服务的配置
高级功能如多租户和计费为闭源

README 快速开始

Bike4Mind

Your bicycle for the mind in the age of AI.

An open-core AI knowledge platform — a multi-model workspace where notebooks, autonomous agents, and a RAG knowledge engine work together to augment any cognitive task.

What is Bike4Mind?

Bike4Mind is an AI-powered knowledge platform that wires a superset of AI technologies together behind one workspace. Bring your own models, keep your data, and let agents do the heavy lifting.

Notebooks — the core workspace: chat, documents, and context in one place, across any model.
Multi-LLM by design — swap freely between Anthropic, OpenAI, Google Gemini, xAI, Ollama (local), and more, plus image models (FLUX, DALL·E 3, Stable Diffusion).
Quest Master — autonomous, multi-step task planning: text and image generation, vision review, web search, math, code, and human-in-the-loop steps, run in parallel.
AI Agents — autonomous ReAct-style agents that carry out complex tasks against your tools and knowledge.
Knowledge Engine — RAG over your documents: smart chunking, vector search, collections, and tagging so agents can reason about what you know.
Artifacts — reusable snippets, documents, and visualizations produced by agents, with a built-in publish-and-share layer.

Try the hosted service at app.bike4mind.com, or self-host the open core on your own infrastructure (below).

Open core

Bike4Mind is open core. The engine is public and self-hostable; the multi-tenant hosted service is our business.

Open (this repo, BUSL-1.1): the agent engine, LLM adapters, CLI, data models, and the self-host path — published as @bike4mind/* packages.
Commercial / closed: operating the multi-tenant hosted service — billing, entitlements, hosted infrastructure — and premium overlays such as Overwatch.

See CONTRIBUTING.md for the exact open/closed boundary.

Quickstart

Hosted (fastest): sign in at app.bike4mind.com — nothing to install.

Self-host: run the open core on your own hardware with Docker — no AWS account or cloud provider required. See the **[Self-Host Quickstart](./

项目描述

The open-core AI workbench — notebooks, agents, RAG, voice, and images across any model: OpenAI, Anthropic, Google, xAI, or local via Ollama/vLLM. BSL 1.1, auto-converting to Apache-2.0 on a two-year clock. Your AI keeps running when theirs doesn't.

bike4mind