[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-83099":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":16,"stars7d":17,"stars30d":17,"stars90d":16,"forks30d":16,"starsTrendScore":18,"compositeScore":19,"rankGlobal":10,"rankLanguage":10,"license":20,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":21,"topics":24,"createdAt":10,"pushedAt":10,"updatedAt":43,"readmeContent":44,"aiSummary":45,"trendingCount":16,"starSnapshotCount":16,"syncStatus":18,"lastSyncTime":46,"discoverSource":47},83099,"continuum","shyftlabs\u002Fcontinuum","shyftlabs","Continuum — the agent runtime by ShyftLabs. Build, orchestrate, ship.","https:\u002F\u002Fdocs.continuum.shyftlabs.io\u002F",null,"Python",73,7,60,12,0,8,2,2.71,"Apache License 2.0",false,"main",true,[25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42],"agent-framework","agentic-ai","ai-agents","ai-orchestration","anthropic","enterprise-ai","human-in-the-loop","kimi-k2","llama","llm","llm-framework","llm-observability","llmops","mcp","multi-agent","openai","qwen","temporal","2026-06-12 02:04:31","\u003Cdiv align=\"center\">\n\n\u003Cpicture>\n  \u003Csource media=\"(prefers-color-scheme: dark)\" srcset=\"docs\u002Fassets\u002Fcontinuum-logo-dark.png\" \u002F>\n  \u003Cimg src=\"docs\u002Fassets\u002Fcontinuum-logo.png\" alt=\"Continuum\" width=\"460\" \u002F>\n\u003C\u002Fpicture>\n\n##### by **[ShyftLabs](https:\u002F\u002Fshyftlabs.io\u002F)**\n\n### The agent runtime for builders who ship.\n\nBuild, run, and deploy reliable AI agents at enterprise scale — multi-LLM routing, persistent memory, MCP-native tools, durable workflows, and full observability, out of the box.\n\n\u003Cbr \u002F>\n\n[![Python 3.13+](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-3.13+-0a0a0a.svg?style=for-the-badge&logo=python&logoColor=white)](https:\u002F\u002Fwww.python.org\u002Fdownloads\u002F)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-Apache_2.0-0a0a0a.svg?style=for-the-badge)](LICENSE)\n[![Version](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fversion-0.2.0-0a0a0a.svg?style=for-the-badge)](https:\u002F\u002Fgithub.com\u002Fshyftlabs\u002Fcontinuum\u002Freleases)\n\n[![CI](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Factions\u002Fworkflow\u002Fstatus\u002Fshyftlabs\u002Fcontinuum\u002Fci.yml?branch=main&label=CI&logo=github)](https:\u002F\u002Fgithub.com\u002Fshyftlabs\u002Fcontinuum\u002Factions\u002Fworkflows\u002Fci.yml)\n[![Docs](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdocs-continuum.shyftlabs.io-blue?logo=readthedocs&logoColor=white)](https:\u002F\u002Fdocs.continuum.shyftlabs.io\u002F)\n[![PRs Welcome](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FPRs-welcome-brightgreen.svg)](CONTRIBUTING.md)\n[![Code of Conduct](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fcode%20of%20conduct-v2.1-ff69b4.svg)](CODE_OF_CONDUCT.md)\n\n[**📖 Documentation**](https:\u002F\u002Fdocs.continuum.shyftlabs.io\u002F) · [**⚡ Quick start**](#-quick-start) · [**⚙️ Configuration**](#️-configuring-continuum) · [**🧩 Components**](#-components) · [**🧪 Examples**](#-examples) · [**🤝 Contributing**](CONTRIBUTING.md)\n\n\u003C\u002Fdiv>\n\n---\n\n**Continuum** is a production-grade Python framework for building, orchestrating, and shipping autonomous AI agents at enterprise scale. It unifies a clean, typed agent core with cost-aware multi-model inference, stateful long- and short-term memory, open standards-based tool calling, durable execution, and end-to-end observability — all behind one small, composable, type-safe API.\n\n## ✨ Features\n\n- 🤖 **Agentic core & orchestration** — a strongly-typed agent primitive with full lifecycle hooks, schema-validated structured outputs, and nine composable multi-agent patterns (sequential, parallel, loop, routing, planning, reflection, debate, scatter, supervised).\n- 🔀 **Smart Inference** — cost-aware inference routing that classifies every request by complexity and dispatches it to the cheapest capable model, with seamless cross-provider failover and zero lock-in.\n- 🧠 **Stateful memory** — persistent semantic long-term recall plus low-latency working memory, with multi-tenant isolation scopes and built-in PII redaction for privacy-by-default agents.\n- 🔌 **Open tool calling** — plug into any standards-based tool ecosystem (Model Context Protocol) across multiple transports, with fine-grained capability scoping, context capture\u002Finjection, and rich generative-UI artifacts.\n- 🔁 **Durable execution** — long-running, crash- and restart-safe agent workflows with human-in-the-loop approval gates and exactly-once guarantees.\n- 🔭 **Full observability** — first-class distributed tracing, token\u002Flatency\u002Ferror telemetry, and one-line function instrumentation for complete run transparency.\n- 🌐 **Model-agnostic** — target frontier or open-weight models through a single model string; swap providers without touching agent code.\n- 🤝 **Multi-agent handoffs** — context-preserving agent-to-agent delegation with history summarization, cycle detection, and depth control.\n- 📡 **Real-time streaming** — token-, tool-, handoff-, and memory-level events streamed the moment they happen.\n- ✅ **Built-in evaluation** — turn live production traces into golden datasets and regression-test agent quality with standard LLM-evaluation metrics.\n\n## 🚀 Quick start\n\n**Requirements:** Python 3.13+ and Docker (for Redis · Milvus\u002FQdrant · Langfuse).\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fshyftlabs\u002Fcontinuum.git\ncd continuum\n\npython3.13 -m venv .venv && source .venv\u002Fbin\u002Factivate\npip install -e .\n\ncp .env.template .env        # add your provider key(s) — see Configuration below\ndocker compose up -d         # Redis · Milvus\u002FQdrant · Langfuse\n```\n\nYour first agent:\n\n```python\nimport asyncio\nfrom orchestrator.agent import BaseAgent, AgentRunner\n\nasync def main():\n    agent = BaseAgent(\n        name=\"hello-agent\",\n        instructions=\"You are a friendly assistant.\",\n        model=\"gpt-4o-mini\",\n    )\n    runner = AgentRunner()\n    response = await runner.run(agent, \"Hi!\")\n    print(response.content)\n\nasyncio.run(main())\n```\n\n`AgentRunner.run()` returns an `AgentResponse` with `content`, `structured_output`, `usage`, `tool_calls`, `run_artifacts`, `latency_ms`, and the full handoff chain. See the [**docs**](https:\u002F\u002Fdocs.continuum.shyftlabs.io\u002F) for streaming, tools\u002FMCP, memory, handoffs, and workflows.\n\n## ⚙️ Configuring Continuum\n\nContinuum is configured through environment variables (copy [`.env.template`](.env.template) → `.env`). Set keys only for the providers and components you use — everything else has sensible defaults. The most common settings:\n\n#### LLM providers & routing\n\n| Variable | Description | Example |\n|---|---|---|\n| `OPENAI_API_KEY` \u002F `ANTHROPIC_API_KEY` \u002F `GEMINI_API_KEY` | Provider API keys — set the one(s) you use | `sk-…` |\n| `DEFAULT_LLM_MODEL` | Default model (`provider\u002Fmodel`, or bare name for OpenAI) | `gemini\u002Fgemini-2.5-flash` |\n| `FALLBACK_LLM_MODEL` | Model used if the default fails | `gpt-4o-mini` |\n| `LLM_ENABLE_FALLBACK` | Automatically fall back on provider errors | `true` |\n| `SMART_LAYER_ENABLED` | Enable cost-aware tier routing (Smart Inference) | `true` |\n\n#### Memory (long-term) & embeddings\n\n| Variable | Description | Example |\n|---|---|---|\n| `MEMORY_ENABLED` | Enable mem0-backed long-term memory | `true` |\n| `VECTOR_STORE_PROVIDER` | Vector store backend | `qdrant` \u002F `milvus` |\n| `EMBEDDER_PROVIDER` \u002F `EMBEDDER_MODEL` | Embedding provider & model | `openai` \u002F `text-embedding-3-small` |\n| `MEMORY_ISOLATION` | Scope of memory isolation | `user` \u002F `agent` \u002F `run` \u002F `shared` |\n\n#### Sessions (short-term)\n\n| Variable | Description | Example |\n|---|---|---|\n| `SESSION_ENABLED` | Enable Redis-backed conversation sessions | `true` |\n| `SESSION_REDIS_HOST` \u002F `SESSION_REDIS_PORT` | Redis connection | `localhost` \u002F `6380` |\n| `SESSION_TTL_SECONDS` | Session lifetime | `172800` |\n\n#### Observability (Langfuse)\n\n| Variable | Description | Example |\n|---|---|---|\n| `LANGFUSE_ENABLED` | Enable tracing | `true` |\n| `LANGFUSE_PUBLIC_KEY` \u002F `LANGFUSE_SECRET_KEY` | Langfuse credentials | `pk-…` \u002F `sk-…` |\n| `LANGFUSE_HOST` | Langfuse endpoint | `http:\u002F\u002Flocalhost:3000` |\n\n#### Temporal (optional, durable workflows)\n\n| Variable | Description | Example |\n|---|---|---|\n| `TEMPORAL_ENABLED` | Enable durable workflow orchestration | `false` |\n| `TEMPORAL_HOST` | Temporal frontend | `localhost:7233` |\n\n> Optional extras: `pip install -e \".[temporal]\"` for Temporal, `\".[eval]\"` for evaluation, `\".[embeddings]\"` for local embeddings. See [`.env.template`](.env.template) for the complete, annotated reference.\n\n## 🧩 Components\n\n| Component | What it does |\n|---|---|\n| **Agents** | `BaseAgent` + `AgentRunner` — config, hooks, structured outputs, ReAct |\n| **Workflows** | Nine multi-agent patterns for chaining, branching, looping, and self-improvement |\n| **Smart Inference** | Request classifier + cost-aware model routing with fallback |\n| **Memory** | mem0 + Qdrant\u002FMilvus (long-term) · Redis (sessions) · multi-tenant scopes |\n| **Tools \u002F MCP** | MCP servers over Stdio\u002FSSE\u002FStreamableHTTP, tool filtering, widget artifacts |\n| **Temporal** | Durable, restart-safe workflows with human-in-the-loop gates |\n| **Observability** | Langfuse traces, metrics, `@observe` decorators |\n| **Evaluation** | Golden datasets + DeepEval \u002F RAGAS metrics |\n\n## 📚 Documentation\n\nFull documentation lives at **[docs.continuum.shyftlabs.io](https:\u002F\u002Fdocs.continuum.shyftlabs.io\u002F)** — guides for building & running agents, Smart Inference, memory, tools\u002FMCP, workflows, handoffs, streaming, evaluation, and the research behind it.\n\nMarkdown sources are also in [`docs\u002F`](docs\u002F) if you prefer reading on GitHub — e.g. [`agent.md`](docs\u002Fagent.md), [`memory.md`](docs\u002Fmemory.md), [`tools.md`](docs\u002Ftools.md), and the integration [`GUIDE.md`](docs\u002FGUIDE.md).\n\n## 🧪 Examples\n\nRunnable demos live under [`playground\u002F`](playground\u002F):\n\n- **`gateway-local-shop`** — an MCP server + agent + chat UI for a pet-shop assistant (end-to-end: server → agent → UI).\n- **`gateway-multi-agent-shop`** — a multi-agent workflow variant with routing and handoffs.\n- **`frontend\u002F`** — the demo web UIs (`assortment`, `commerce-chat`).\n\n## 🤝 Contributing\n\nContributions are welcome! Please read [`CONTRIBUTING.md`](CONTRIBUTING.md) for the branch model, Conventional Commits, DCO sign-off, and local setup. By participating you agree to our [Code of Conduct](CODE_OF_CONDUCT.md).\n\n- 🐛 **Bugs & features** — use the [issue templates](.github\u002FISSUE_TEMPLATE)\n- 💬 **Questions & ideas** — [GitHub Discussions](https:\u002F\u002Fgithub.com\u002Fshyftlabs\u002Fcontinuum\u002Fdiscussions)\n- 🔒 **Security** — report privately via [`SECURITY.md`](SECURITY.md), never a public issue\n\n## 📄 License\n\nLicensed under the [Apache License, Version 2.0](LICENSE). Copyright © 2025–2026 [ShyftLabs Inc.](https:\u002F\u002Fshyftlabs.io\u002F)\n\nFor commercial \u002F enterprise inquiries — SLAs, indemnification, hosted offerings, custom features — contact **[continuum@shyftlabs.io](mailto:continuum@shyftlabs.io)**.\n\n\u003Cdiv align=\"center\">\n\u003Cbr \u002F>\n\u003Csub>Built with ❤️ by \u003Ca href=\"https:\u002F\u002Fshyftlabs.io\u002F\">ShyftLabs\u003C\u002Fa> · \u003Ca href=\"mailto:continuum@shyftlabs.io\">continuum@shyftlabs.io\u003C\u002Fa>\u003C\u002Fsub>\n\u003C\u002Fdiv>\n","Continuum 是一个用于构建、编排和部署企业级AI代理的Python框架。其核心功能包括强类型代理内核支持完整的生命周期钩子及结构化输出，智能推理路由能够根据请求复杂度选择最经济的模型进行处理，并具备持久化记忆能力以实现长期和短期的数据存储与访问。此外，它还支持开放标准工具调用，允许跨多个传输协议接入工具生态系统。适用于需要在大规模生产环境中运行可靠且可观察的多代理系统场景，如企业级AI解决方案开发、多语言模型协同工作等。","2026-06-11 04:10:06","CREATED_QUERY"]