[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-70763":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":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":25,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":46,"readmeContent":47,"aiSummary":48,"trendingCount":16,"starSnapshotCount":16,"syncStatus":49,"lastSyncTime":50,"discoverSource":51},70763,"txtai","neuml\u002Ftxtai","neuml","💡 All-in-one AI framework for semantic search, LLM orchestration and language model workflows","https:\u002F\u002Fneuml.github.io\u002Ftxtai",null,"Python",12649,832,112,9,0,12,27,169,36,43.76,"Apache License 2.0",false,"master",true,[27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,5,44,45],"agents","ai","ai-agents","embeddings","information-retrieval","language-model","large-language-models","llm","nlp","python","rag","retrieval-augmented-generation","search","search-engine","semantic-search","sentence-embeddings","transformers","vector-database","vector-search","2026-06-12 02:02:43","\u003Cp align=\"center\">\n    \u003Cimg src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Flogo.png\"\u002F>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n    \u003Cb>All-in-one AI framework\u003C\u002Fb>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Freleases\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Frelease\u002Fneuml\u002Ftxtai.svg?style=flat&color=success\" alt=\"Version\"\u002F>\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flast-commit\u002Fneuml\u002Ftxtai.svg?style=flat&color=blue\" alt=\"GitHub last commit\"\u002F>\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fissues\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Fissues\u002Fneuml\u002Ftxtai.svg?style=flat&color=success\" alt=\"GitHub issues\"\u002F>\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fjoin.slack.com\u002Ft\u002Ftxtai\u002Fshared_invite\u002Fzt-37c1zfijp-Y57wMty6YOx_hyIHEQvQJA\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fslack-join-blue?style=flat&logo=slack&logocolor=white\" alt=\"Join Slack\"\u002F>\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Factions?query=workflow%3Abuild\">\n        \u003Cimg src=\"https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fworkflows\u002Fbuild\u002Fbadge.svg\" alt=\"Build Status\"\u002F>\n    \u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fcoveralls.io\u002Fgithub\u002Fneuml\u002Ftxtai?branch=master\">\n        \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002FcoverallsCoverage\u002Fgithub\u002Fneuml\u002Ftxtai\" alt=\"Coverage Status\">\n    \u003C\u002Fa>\n\u003C\u002Fp>\n\ntxtai is an all-in-one AI framework for semantic search, LLM orchestration and language model workflows.\n\n![architecture](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Farchitecture.png#gh-light-mode-only)\n![architecture](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Farchitecture-dark.png#gh-dark-mode-only)\n\nThe key component of txtai is an embeddings database, which is a union of vector indexes (sparse and dense), graph networks and relational databases.\n\nThis foundation enables vector search and\u002For serves as a powerful knowledge source for large language model (LLM) applications.\n\nBuild autonomous agents, retrieval augmented generation (RAG) processes, multi-model workflows and more.\n\nSummary of txtai features:\n\n- 🔎 Vector search with SQL, object storage, topic modeling, graph analysis and multimodal indexing\n- 📄 Create embeddings for text, documents, audio, images and video\n- 💡 Pipelines powered by language models that run LLM prompts, question-answering, labeling, transcription, translation, summarization and more\n- ↪️️ Workflows to join pipelines together and aggregate business logic. txtai processes can be simple microservices or multi-model workflows.\n- 🤖 Agents that intelligently connect embeddings, pipelines, workflows and other agents together to autonomously solve complex problems\n- ⚙️ Web and Model Context Protocol (MCP) APIs. Bindings available for [JavaScript](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai.js), [Java](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai.java), [Rust](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai.rs) and [Go](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai.go).\n- 🔋 Batteries included with defaults to get up and running fast\n- ☁️ Run local or scale out with container orchestration\n\ntxtai is built with Python 3.10+, [Hugging Face Transformers](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Ftransformers), [Sentence Transformers](https:\u002F\u002Fgithub.com\u002FUKPLab\u002Fsentence-transformers) and [FastAPI](https:\u002F\u002Fgithub.com\u002Ftiangolo\u002Ffastapi). txtai is open-source under an Apache 2.0 license.\n\n> [!NOTE]\n>\n> [NeuML](https:\u002F\u002Fneuml.com) is the company behind txtai and we provide AI consulting services around our stack. [Schedule a meeting](https:\u002F\u002Fcal.com\u002Fneuml\u002Fintro) or [send a message](mailto:info@neuml.com) to learn more.\n>\n> We're also building an easy and secure way to run hosted txtai applications with [txtai.cloud](https:\u002F\u002Ftxtai.cloud).\n\n## Why txtai?\n\n![why](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fwhy.png#gh-light-mode-only)\n![why](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fwhy-dark.png#gh-dark-mode-only)\n\nNew vector databases, LLM frameworks and everything in between are sprouting up daily. Why build with txtai?\n\n- Up and running in minutes with [pip](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Finstall\u002F) or [Docker](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fcloud\u002F)\n```python\n# Get started in a couple lines\nimport txtai\n\nembeddings = txtai.Embeddings()\nembeddings.index([\"Correct\", \"Not what we hoped\"])\nembeddings.search(\"positive\", 1)\n#[(0, 0.29862046241760254)]\n```\n- Built-in API makes it easy to develop applications using your programming language of choice\n```yaml\n# app.yml\nembeddings:\n    path: sentence-transformers\u002Fall-MiniLM-L6-v2\n```\n```bash\nCONFIG=app.yml uvicorn \"txtai.api:app\"\ncurl -X GET \"http:\u002F\u002Flocalhost:8000\u002Fsearch?query=positive\"\n```\n- Run local - no need to ship data off to disparate remote services\n- Work with micromodels all the way up to large language models (LLMs)\n- Low footprint - install additional dependencies and scale up when needed\n- [Learn by example](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fexamples) - notebooks cover all available functionality\n\n## Use Cases\n\nThe following sections introduce common txtai use cases. A comprehensive set of over 70 [example notebooks and applications](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fexamples) are also available.\n\n### Semantic Search\n\nBuild semantic\u002Fsimilarity\u002Fvector\u002Fneural search applications.\n\n![demo](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdemo.gif)\n\nTraditional search systems use keywords to find data. Semantic search has an understanding of natural language and identifies results that have the same meaning, not necessarily the same keywords.\n\n![search](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fsearch.png#gh-light-mode-only)\n![search](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fsearch-dark.png#gh-dark-mode-only)\n\nGet started with the following examples.\n\n| Notebook  | Description  |       |\n|:----------|:-------------|------:|\n| [Introducing txtai](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F01_Introducing_txtai.ipynb) [▶️](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=SIezMnVdmMs) | Overview of the functionality provided by txtai | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F01_Introducing_txtai.ipynb) |\n| [Similarity search with images](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F13_Similarity_search_with_images.ipynb) | Embed images and text into the same space for search | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F13_Similarity_search_with_images.ipynb) |\n| [Build a QA database](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F34_Build_a_QA_database.ipynb) | Question matching with semantic search | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F34_Build_a_QA_database.ipynb) |\n| [Semantic Graphs](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F38_Introducing_the_Semantic_Graph.ipynb) | Explore topics, data connectivity and run network analysis| [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F38_Introducing_the_Semantic_Graph.ipynb) |\n\n### LLM Orchestration\n\nAutonomous agents, retrieval augmented generation (RAG), chat with your data, pipelines and workflows that interface with large language models (LLMs).\n\n![llm](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fllm.png)\n\nSee below to learn more.\n\n| Notebook  | Description  |       |\n|:----------|:-------------|------:|\n| [Prompt templates and task chains](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F44_Prompt_templates_and_task_chains.ipynb) | Build model prompts and connect tasks together with workflows | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F44_Prompt_templates_and_task_chains.ipynb) |\n| [Integrate LLM frameworks](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F53_Integrate_LLM_Frameworks.ipynb) | Integrate llama.cpp, LiteLLM and custom generation frameworks | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F53_Integrate_LLM_Frameworks.ipynb) |\n| [Build knowledge graphs with LLMs](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F57_Build_knowledge_graphs_with_LLM_driven_entity_extraction.ipynb) | Build knowledge graphs with LLM-driven entity extraction | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F57_Build_knowledge_graphs_with_LLM_driven_entity_extraction.ipynb) |\n| [Parsing the stars with txtai](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F72_Parsing_the_stars_with_txtai.ipynb) | Explore an astronomical knowledge graph of known stars, planets, galaxies | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F72_Parsing_the_stars_with_txtai.ipynb) |\n\n#### Agents\n\nAgents connect embeddings, pipelines, workflows and other agents together to autonomously solve complex problems.\n\n![agent](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fagent.png)\n\ntxtai agents are built on top of the [smolagents](https:\u002F\u002Fgithub.com\u002Fhuggingface\u002Fsmolagents) framework. This supports all LLMs txtai supports (Hugging Face, llama.cpp, OpenAI \u002F Claude \u002F AWS Bedrock via LiteLLM). Agent prompting with [`agents.md`](https:\u002F\u002Fgithub.com\u002Fagentsmd\u002Fagents.md) and [`skill.md`](https:\u002F\u002Fagentskills.io\u002Fspecification) are also supported.\n\nCheck out this [Agent Quickstart Example](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002Fagent_quickstart.py). Additional examples are listed below.\n\n| Notebook  | Description  |       |\n|:----------|:-------------|------:|\n| [Granting autonomy to agents](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F69_Granting_autonomy_to_agents.ipynb) | Agents that iteratively solve problems as they see fit | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F69_Granting_autonomy_to_agents.ipynb) |\n| [TxtAI got skills](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F83_TxtAI_got_skills.ipynb) | Integrate skill.md files with your agent | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F83_TxtAI_got_skills.ipynb) |\n| [Agent Tools](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F84_Agent_Tools.ipynb) [▶️](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=RDNaFXQy3GQ) | Learn about the txtai agent toolkit | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F84_Agent_Tools.ipynb) |\n| [Analyzing LinkedIn Company Posts with Graphs and Agents](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F71_Analyzing_LinkedIn_Company_Posts_with_Graphs_and_Agents.ipynb) | Exploring how to improve social media engagement with AI | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F71_Analyzing_LinkedIn_Company_Posts_with_Graphs_and_Agents.ipynb) |\n\n#### Retrieval augmented generation\n\nRetrieval augmented generation (RAG) reduces the risk of LLM hallucinations by constraining the output with a knowledge base as context. RAG is commonly used to \"chat with your data\".\n\n![rag](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Frag.png#gh-light-mode-only)\n![rag](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Frag-dark.png#gh-dark-mode-only)\n\nCheck out this [RAG Quickstart Example](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002Frag_quickstart.py). Additional examples are listed below.\n\n| Notebook  | Description  |       |\n|:----------|:-------------|------:|\n| [Build RAG pipelines with txtai](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F52_Build_RAG_pipelines_with_txtai.ipynb) [▶️](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=t_OeAc8NVfQ) | Guide on retrieval augmented generation including how to create citations | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F52_Build_RAG_pipelines_with_txtai.ipynb) |\n| [RAG is more than Vector Search](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F79_RAG_is_more_than_Vector_Search.ipynb) | Context retrieval via Web, SQL and other sources | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F79_RAG_is_more_than_Vector_Search.ipynb) |\n| [GraphRAG with Wikipedia and GPT OSS](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F77_GraphRAG_with_Wikipedia_and_GPT_OSS.ipynb) | Deep graph search powered RAG | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F77_GraphRAG_with_Wikipedia_and_GPT_OSS.ipynb) |\n| [Speech to Speech RAG](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F65_Speech_to_Speech_RAG.ipynb) [▶️](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=tH8QWwkVMKA) | Full cycle speech to speech workflow with RAG | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F65_Speech_to_Speech_RAG.ipynb) |\n\n### Language Model Workflows\n\nLanguage model workflows, also known as semantic workflows, connect language models together to build intelligent applications.\n\n![flows](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fflows.png#gh-light-mode-only)\n![flows](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fflows-dark.png#gh-dark-mode-only)\n\nWhile LLMs are powerful, there are plenty of smaller, more specialized models that work better and faster for specific tasks. This includes models for extractive question-answering, automatic summarization, text-to-speech, transcription and translation.\n\nCheck out this [Workflow Quickstart Example](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002Fworkflow_quickstart.py). Additional examples are listed below.\n\n| Notebook  | Description  |       |\n|:----------|:-------------|------:|\n| [Run pipeline workflows](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F14_Run_pipeline_workflows.ipynb) [▶️](https:\u002F\u002Fwww.youtube.com\u002Fwatch?v=UBMPDCn1gEU) | Simple yet powerful constructs to efficiently process data | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F14_Run_pipeline_workflows.ipynb) |\n| [Building abstractive text summaries](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F09_Building_abstractive_text_summaries.ipynb) | Run abstractive text summarization | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F09_Building_abstractive_text_summaries.ipynb) |\n| [Transcribe audio to text](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F11_Transcribe_audio_to_text.ipynb) | Convert audio files to text | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F11_Transcribe_audio_to_text.ipynb) |\n| [Translate text between languages](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F12_Translate_text_between_languages.ipynb) | Streamline machine translation and language detection | [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuml\u002Ftxtai\u002Fblob\u002Fmaster\u002Fexamples\u002F12_Translate_text_between_languages.ipynb) |\n\n## Installation\n\n![install](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Finstall.png#gh-light-mode-only)\n![install](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Finstall-dark.png#gh-dark-mode-only)\n\nThe easiest way to install is via pip and PyPI\n\n```\npip install txtai\n```\n\nPython 3.10+ is supported. Using a Python [virtual environment](https:\u002F\u002Fdocs.python.org\u002F3\u002Flibrary\u002Fvenv.html) is recommended.\n\nSee the detailed [install instructions](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Finstall) for more information covering [optional dependencies](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Finstall\u002F#optional-dependencies), [environment specific prerequisites](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Finstall\u002F#environment-specific-prerequisites), [installing from source](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Finstall\u002F#install-from-source), [conda support](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Finstall\u002F#conda), [lightweight minimal installation](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Finstall\u002F#minimal-install) and how to [run with containers](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fcloud).\n\n## Model guide\n\n![models](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Fmodels.png)\n\nSee the table below for the current recommended models. These models all allow commercial use and offer a blend of speed and performance.\n\n| Component                                                                     | Model(s)                                                                 |\n| ----------------------------------------------------------------------------- | ------------------------------------------------------------------------ |\n| [Embeddings](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fembeddings)                        | [all-MiniLM-L6-v2](https:\u002F\u002Fhf.co\u002Fsentence-transformers\u002Fall-MiniLM-L6-v2) | \n| [Image Captions](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Fimage\u002Fcaption)        | [BLIP](https:\u002F\u002Fhf.co\u002FSalesforce\u002Fblip-image-captioning-base)              |\n| [Labels - Zero Shot](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Ftext\u002Flabels)      | [DeBERTa v3 Zeroshot](https:\u002F\u002Fhf.co\u002FMoritzLaurer\u002Fdeberta-v3-base-zeroshot-v2.0-c) |\n| [Labels - Fixed](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Ftext\u002Flabels)          | Fine-tune with [training pipeline](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Ftrain\u002Ftrainer)          |\n| [Large Language Model (LLM)](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Ftext\u002Fllm) | [Gemma 4 31B](https:\u002F\u002Fhf.co\u002Fgoogle\u002Fgemma-4-31B) |\n| [Summarization](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Ftext\u002Fsummary)          | [DistilBART](https:\u002F\u002Fhf.co\u002Fsshleifer\u002Fdistilbart-cnn-12-6)                |\n| [Text-to-Speech](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Faudio\u002Ftexttospeech)   | [ESPnet JETS](https:\u002F\u002Fhf.co\u002FNeuML\u002Fljspeech-jets-onnx)                    |\n| [Transcription](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Faudio\u002Ftranscription)   | [Whisper](https:\u002F\u002Fhf.co\u002Fopenai\u002Fwhisper-base)                             | \n| [Translation](https:\u002F\u002Fneuml.github.io\u002Ftxtai\u002Fpipeline\u002Ftext\u002Ftranslation)        | [OPUS Model Series](https:\u002F\u002Fhf.co\u002FHelsinki-NLP)                          |\n\nModels can be loaded as either a path from the Hugging Face Hub or a local directory. Model paths are optional, defaults are loaded when not specified. For tasks with no recommended model, txtai uses the default models as shown in the Hugging Face Tasks guide.\n\nSee the following links to learn more.\n\n- [Hugging Face Tasks](https:\u002F\u002Fhf.co\u002Ftasks)\n- [Hugging Face Model Hub](https:\u002F\u002Fhf.co\u002Fmodels)\n- [Embeddings Leaderboard](https:\u002F\u002Fhf.co\u002Fspaces\u002Fmteb\u002Fleaderboard)\n- [LLM Leaderboard](https:\u002F\u002Fhf.co\u002Fspaces\u002Flmarena-ai\u002Farena-leaderboard)\n\n## Powered by txtai\n\nThe following applications are powered by txtai.\n\n![apps](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fapps.jpg)\n\n| Application  | Description  |\n|:------------ |:-------------|\n| [rag](https:\u002F\u002Fgithub.com\u002Fneuml\u002Frag) | Retrieval Augmented Generation (RAG) application |\n| [ncoder](https:\u002F\u002Fgithub.com\u002Fneuml\u002Fncoder) | Open-Source AI coding agent |\n| [paperai](https:\u002F\u002Fgithub.com\u002Fneuml\u002Fpaperai) | AI for medical and scientific papers |\n| [annotateai](https:\u002F\u002Fgithub.com\u002Fneuml\u002Fannotateai) | Automatically annotate papers with LLMs |\n\nIn addition to this list, there are also many other [open-source projects](https:\u002F\u002Fgithub.com\u002Fneuml\u002Ftxtai\u002Fnetwork\u002Fdependents), [published research](https:\u002F\u002Fscholar.google.com\u002Fscholar?q=txtai&hl=en&as_ylo=2022) and closed proprietary\u002Fcommercial projects that have built on txtai in production.\n\n## Further Reading\n\n![further](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Ffurther.png#gh-light-mode-only)\n![further](https:\u002F\u002Fraw.githubusercontent.com\u002Fneuml\u002Ftxtai\u002Fmaster\u002Fdocs\u002Fimages\u002Ffurther-ghdark.png#gh-dark-mode-only)\n\n- [Introducing txtai, the all-in-one AI framework](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fintroducing-txtai-the-all-in-one-ai-framework-0660ecfc39d7)\n- [txtai: An All-in-One AI Framework for Semantic Search and LLM Workflows](https:\u002F\u002Fgithub.com\u002Fneuml\u002Fpapers\u002Fblob\u002Fmaster\u002Ftxtai\u002Ftxtai.pdf)\n- [Tutorial series on Hashnode](https:\u002F\u002Fneuml.hashnode.dev\u002Fseries\u002Ftxtai-tutorial) | [dev.to](https:\u002F\u002Fdev.to\u002Fneuml\u002Ftutorial-series-on-txtai-ibg)\n- [What's new in txtai 9.0](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fwhats-new-in-txtai-9-0-d522bb150afa) | [8.0](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fwhats-new-in-txtai-8-0-2d7d0ab4506b) | [7.0](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fwhats-new-in-txtai-7-0-855ad6a55440) | [6.0](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fwhats-new-in-txtai-6-0-7d93eeedf804) | [5.0](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fwhats-new-in-txtai-5-0-e5c75a13b101) | [4.0](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fwhats-new-in-txtai-4-0-bbc3a65c3d1c)\n- [Getting started with semantic search](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fgetting-started-with-semantic-search-a9fd9d8a48cf) | [workflows](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fgetting-started-with-semantic-workflows-2fefda6165d9) | [rag](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fgetting-started-with-rag-9a0cca75f748)\n- [Running txtai at scale](https:\u002F\u002Fmedium.com\u002Fneuml\u002Frunning-at-scale-with-txtai-71196cdd99f9)\n- [Vector search & RAG Landscape: A review with txtai](https:\u002F\u002Fmedium.com\u002Fneuml\u002Fvector-search-rag-landscape-a-review-with-txtai-a7f37ad0e187)\n\n## Documentation\n\n[Full documentation on txtai](https:\u002F\u002Fneuml.github.io\u002Ftxtai) including configuration settings for embeddings, pipelines, workflows, API and a FAQ with common questions\u002Fissues is available.\n\n## Contributing\n\nFor those who would like to contribute to txtai, please see [this guide](https:\u002F\u002Fgithub.com\u002Fneuml\u002F.github\u002Fblob\u002Fmaster\u002FCONTRIBUTING.md).\n","txtai 是一个集成了语义搜索、大语言模型编排和语言模型工作流的全能AI框架。其核心功能包括基于嵌入数据库的向量搜索，支持文本、文档、音频、图像和视频等多模态数据的处理，以及通过流水线和工作流实现复杂的业务逻辑。该框架适合需要高效信息检索、构建自主代理、执行检索增强生成（RAG）过程或开发涉及多个模型交互的应用场景。此外，txtai 提供了丰富的API接口支持，包括Web和MCP协议，并且有多种编程语言的绑定库可用，使得开发者能够快速上手并灵活地根据需求定制解决方案。",2,"2026-06-11 03:34:04","high_star"]