{"id":33345775,"url":"https://github.com/adrmurbau/ai-agent","last_synced_at":"2026-04-11T00:59:27.580Z","repository":{"id":325341998,"uuid":"1100807364","full_name":"adrmurbau/ai-agent","owner":"adrmurbau","description":"AI agent that answers questions about your PDF/TXT using free RAG (TF-IDF + FLAN-T5) with FastAPI, Gradio and Docker.","archived":false,"fork":false,"pushed_at":"2025-11-20T19:54:06.000Z","size":10,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-11-20T21:17:08.659Z","etag":null,"topics":["ai-agent","docker","fastapi","gradio","python","rag"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/adrmurbau.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":null,"agents":null,"dco":null,"cla":null}},"created_at":"2025-11-20T19:34:06.000Z","updated_at":"2025-11-20T19:54:09.000Z","dependencies_parsed_at":null,"dependency_job_id":null,"html_url":"https://github.com/adrmurbau/ai-agent","commit_stats":null,"previous_names":["adrmurbau/ai-agent"],"tags_count":null,"template":false,"template_full_name":null,"purl":"pkg:github/adrmurbau/ai-agent","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adrmurbau%2Fai-agent","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adrmurbau%2Fai-agent/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adrmurbau%2Fai-agent/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adrmurbau%2Fai-agent/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/adrmurbau","download_url":"https://codeload.github.com/adrmurbau/ai-agent/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/adrmurbau%2Fai-agent/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":285737153,"owners_count":27223129,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","status":"online","status_checked_at":"2025-11-22T02:00:05.934Z","response_time":64,"last_error":null,"robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":true,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["ai-agent","docker","fastapi","gradio","python","rag"],"created_at":"2025-11-22T05:00:31.575Z","updated_at":"2025-11-22T05:01:17.288Z","avatar_url":"https://github.com/adrmurbau.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🧠 Free AI Agent – RAG + FLAN-T5 (100% Free, No API Keys)\n\nAgente de IA que responde preguntas sobre tus documentos **PDF / TXT** usando un pipeline completamente gratuito:\n\n- 🔎 **RAG ligero** con recuperación por **TF-IDF** (scikit-learn)  \n- 🤖 **FLAN-T5** como modelo generador (Transformers, CPU)  \n- ⚙️ **Backend en FastAPI** completamente dockerizado  \n- 🖥️ **Interfaz gráfica en Gradio**  \n- 🧪 Funciona **offline**, sin tokens ni APIs externas  \n\nEste proyecto está pensado como un **MVP limpio y entendible** para portfolio: enseña cómo construir un agente RAG desde cero usando solo herramientas gratuitas.\n\n---\n\n## ✨ Funcionalidades principales\n\n- Ingesta de uno o varios documentos (TXT / PDF con texto)\n- Chunking automático del contenido\n- Creación de índice TF-IDF para recuperar contexto relevante\n- Respuestas generadas usando solo la información del documento\n- API REST (`/ingest` y `/ask`)\n- UI visual en Gradio para uso cómodo\n- Docker para despliegue local reproducible\n\n---\n\n## 📦 Backend (FastAPI + Docker)\n\nDesde la **raíz** del proyecto:\n\n```bash\ndocker build -t ai-agent-backend ./backend\ndocker run --rm -p 8000:8000 ai-agent-backend\n```\n\nEl backend estará disponible en:\n\n👉 http://localhost:8000/docs\n\nDesde ahí puedes probar los endpoints /ingest y /ask.\n\n## 🖥️ UI opcional (Gradio)\nEn otra terminal:\n\n```bash\npy -m pip install -r backend/requirements.txt\npy app.py\n```\n\nLa interfaz estará disponible en:\n\n👉 http://localhost:7860\n\n## 🧪 Ejemplo de uso con cURL\n1) Documento de prueba\n```bash\necho \"El aprendizaje supervisado usa pares entrada-salida.\" \u003e sample_docs/doc1.txt\n```\n2) Ingesta:\n```bash\ncurl -X POST \"http://localhost:8000/ingest\" -F \"files=@sample_docs/doc1.txt\"\n```\n3) Pregunta:\n```bash\ncurl -X POST \"http://localhost:8000/ask\" -F \"question=¿De qué trata el documento?\" -F \"k=3\"\n```\nRespuesta típica:\n```json\n{\"answer\": \"El documento trata sobre el aprendizaje supervisado y el uso de pares entrada-salida.\"}\n```\n---\n## 📂 Estructura del proyecto\n```bash\n/\n├── backend/\n│   ├── agent.py           # Lógica del RAG + LLM\n│   ├── main.py            # FastAPI\n│   ├── requirements.txt\n│   └── Dockerfile\n├── app.py                 # UI con Gradio\n├── sample_docs/           # Documentos de ejemplo\n├── screenshots/            # (Opcional) Capturas para el README\n└── README.md\n```\n---\n## 🧱 Arquitectura \n1. Ingesta\nLee PDFs/TXTs → limpia texto → genera chunks → construye TF-IDF.\n\n2. Consulta\n* Recupera los 𝑘 chunks más similares mediante coseno.\n* Construye un prompt con el contexto recuperado.\n* Genera respuesta usando FLAN-T5 en CPU.\n\n3.Entrega\n* API REST vía FastAPI\n* UI opcional vía Gradio\n---\n## 🔧 Tecnologías usadas\n* **Python 3.11**\n* **FastAPI**\n* **Transformers (Hugging Face)**\n* **scikit-learn**\n* **pypdf**\n* **Gradio**\n* **Docker**\n---\n## 🚀 Mejoras futuras (roadmap)\n* Añadir stopwords en español al TF-IDF\n* Resaltar el pasaje usado en la respuesta\n* Soporte para CSV/JSON\n* Guardar el índice para no re-ingestar\n* Demo gratuita en Hugging Face Spaces\n---\n## 📜 Licencia\nMIT License – libre para usar, modificar y ampliar.\n\n## 👤 Autor\nAdrián Muriel Bautista\n- GitHub: https://github.com/adrmurbau\n- LinkedIn: https://www.linkedin.com/in/adrmurbau\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fadrmurbau%2Fai-agent","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fadrmurbau%2Fai-agent","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fadrmurbau%2Fai-agent/lists"}