{"id":29031624,"url":"https://github.com/mittapallynitin/podcastai","last_synced_at":"2026-05-04T01:32:34.550Z","repository":{"id":301238411,"uuid":"1007459547","full_name":"mittapallynitin/PodcastAI","owner":"mittapallynitin","description":"Podcast AI backend built with FastAPI, powered by Mistral for LLM summarization, and MCP.","archived":false,"fork":false,"pushed_at":"2025-06-25T20:46:56.000Z","size":11,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":0,"default_branch":"main","last_synced_at":"2025-06-25T21:34:12.049Z","etag":null,"topics":["fastapi","llms","mcp","mistral-ocr","openai","pydantic","python"],"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/mittapallynitin.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}},"created_at":"2025-06-24T03:19:48.000Z","updated_at":"2025-06-25T20:47:00.000Z","dependencies_parsed_at":"2025-06-25T21:34:14.187Z","dependency_job_id":"912bf145-1154-4835-8d54-9fe3c2cada69","html_url":"https://github.com/mittapallynitin/PodcastAI","commit_stats":null,"previous_names":["mittapallynitin/podcastai"],"tags_count":0,"template":false,"template_full_name":null,"purl":"pkg:github/mittapallynitin/PodcastAI","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mittapallynitin%2FPodcastAI","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mittapallynitin%2FPodcastAI/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mittapallynitin%2FPodcastAI/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mittapallynitin%2FPodcastAI/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/mittapallynitin","download_url":"https://codeload.github.com/mittapallynitin/PodcastAI/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/mittapallynitin%2FPodcastAI/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":262044446,"owners_count":23249749,"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","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":["fastapi","llms","mcp","mistral-ocr","openai","pydantic","python"],"created_at":"2025-06-26T10:04:48.453Z","updated_at":"2026-05-04T01:32:34.519Z","avatar_url":"https://github.com/mittapallynitin.png","language":"Python","funding_links":[],"categories":[],"sub_categories":[],"readme":"# 🎙️ Podcast AI\n\n**Podcast AI** is an AI-driven platform that converts research papers (e.g., from ArXiv) into podcast-like narratives. Key features include secure user authentication, automated PDF parsing, script generation via LLM, and dynamic prompt orchestration through MCP.\n\n---\n\n## 🔍 Tech Overview\n\n- **Backend**: FastAPI (Python)  \n- **Auth**: JWT for secure login/signup flows  \n- **PDF → Markdown**: Extract content using libraries like PyMuPDF or PDFMiner  \n- **LLM Integration**: Use *Mistral* for content summarization and markdown conversion  \n- **Script Generation**: Custom logic to create engaging, conversational narration  \n- **Prompts Management**: Prompts are stored and fetched remotely using the **Model Context Protocol (MCP)**, enabling version control and easy updates \n\n---\n\n## ⚙️ Why MCP?\n\nThe *Model Context Protocol (MCP)* is an **open standard** introduced by Anthropic in November 2024. It allows LLM-powered apps to:\n1. Discover available tools or prompts  \n2. Fetch structured prompt templates via JSON‑RPC  \n3. Maintain modular and updateable prompt logic separate from code  \n[oai_citation: en.wikipedia.org](https://en.wikipedia.org/wiki/Model_Context_Protocol?utm_source=chatgpt.com)\n[oai_citation: medium.com](https://medium.com/ai-cloud-lab/model-context-protocol-mcp-with-ollama-a-full-deep-dive-working-code-part-1-81a3bb6d16b3?utm_source=chatgpt.com) \n[oai_citation: blog.miloslavhomer.cz](https://blog.miloslavhomer.cz/p/tools-for-mistral-model-context-protocol?utm_source=chatgpt.com)\n\nMCP is widely adopted by OpenAI, Google DeepMind, Microsoft, and many others as the “USB‑C for AI apps”  \n\n---\n\n## 🏗️ Architecture\n\n```\nUser → FastAPI Endpoints → PDF Fetcher → Markdown Extractor →\nMistral LLM → Script Generator → Output 🎙️\n↑\nPrompts from MCP Server\n```\n\n---\n\n## ✅ Feature List\n\n- **JWT-Based Authentication**  \n- **Paper Ingestion**: Submit ArXiv URLs or PDFs  \n- **PDF → Markdown Extraction**  (Mistral OCR)\n- **Mistral-Powered LLM** summarizing markdown into scripts  \n- **Prompt Orchestration** with MCP server – remote fetch and version control  \n- **Custom Narration Styles** – default presets + user-defined options  \n\n---\n\n## 🧑‍💻 Setup \u0026 Run\n\n```bash\ngit clone https://github.com/…/PodcastAI.git\ncd PodcastAI\n\npython -m venv venv\nsource venv/bin/activate\npip install -r requirements.txt\n\nuvicorn app.main:app --reload\n```\n\n## Backend\n\n### Authentication flow\n\n```bash\nPOST /auth/login\n{\n  \"username\": \"your_user\",\n  \"password\": \"secure_pass\"\n}\n# → { \"access_token\": \"jwt-token\" }\n\nGET /papers?url=https://arxiv.org/abs/…\nAuthorization: Bearer \u003cjwt-token\u003e\n```\n## 🧠 LLM + MCP Prompting\n- MCP Client in your backend auto-discovers prompt templates.\n- Makes JSON-RPC call to MCP server to get best practice and latest prompts.\n- Sends paper’s markdown + prompt to Mistral to generate natural-language script.\n- MCP centralizes prompt updates—improve narrator style without backend changes\n\n## 📦 Sample Request \u0026 Response\n```bash\nPOST /papers/request\nContent-Type: application/json\nAuthorization: Bearer \u003ctoken\u003e\n\n{\n  \"arxiv_url\": \"https://arxiv.org/abs/1706.03762\",\n  \"style_id\": \"concise_explainer\"\n}\n---\n200 OK\n{\n  \"script\": \"In this episode, we explore the Transformer architecture introduced in 2017...\"\n}\n```\n\n### 📌 Roadmap\n- 🗣️ TTS Integration (e.g., ElevenLabs, Bark)\n- 🔁 Job orchestration: background tasks with Celery or RQ\n- 🧑‍🎓 Frontend/UI with live progress (FastAPI + WebSockets/Streamlit)\n- 🎧 Podcast Publishing: export to RSS, audio stores","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmittapallynitin%2Fpodcastai","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmittapallynitin%2Fpodcastai","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmittapallynitin%2Fpodcastai/lists"}