Symfony AI Knowledgebase & Chat. 251217
Budget: $250 – $750 USD
I’m building an AI-powered knowledgebase with a chat interface and I need a skilled Symfony developer to take it from concept to working product. The core idea is simple: I bulk-upload PDF documents, tag them as either “competitor intel” or “our platform docs,” and then query everything through a ChatGPT-style UI—minus the usual front-end memory leaks.
What I already know I want
• Backend in PHP 8.x on the latest stable Symfony.
• Vector-store or similar approach to index the PDFs so answers stay grounded in the source text.
• A chat-based front-end that streams responses smoothly without exhausting browser RAM.
• First integrations with OpenAI’s GPT models, but the code should make it easy to swap in Google’s Gemini and Anthropic’s Claude via their APIs.
• A reliable bulk upload flow (drag-and-drop or folder select) with progress feedback and graceful error handling.
Deliverables
1. Symfony project with database schema, document ingestor, and embeddings pipeline.
2. Responsive chat UI wired to selected LLM endpoint(s).
3. Clear README covering local setup, environment variables for model keys, and how to add a new model provider.
4. Short test script or Postman collection that shows the PDF ingestion and chat query working end-to-end.
5. Committed to my github repo.
Acceptance criteria
• I can upload 10+ PDFs at once without a timeout.
• Chat replies cite their source pages and stay relevant to the chosen knowledge domain.
• Switching from ChatGPT API to Gemini API or Claude API is a UI toggle, not a rewrite.
• Browser memory stays stable during a 1000-turn conversation.
---
In your bid, post how you would structure the project and any libraries you’d reach for (Symfony Messenger, Mercure, Pinecone, etc.) or I will assume you are a bot.
What I already know I want
• Backend in PHP 8.x on the latest stable Symfony.
• Vector-store or similar approach to index the PDFs so answers stay grounded in the source text.
• A chat-based front-end that streams responses smoothly without exhausting browser RAM.
• First integrations with OpenAI’s GPT models, but the code should make it easy to swap in Google’s Gemini and Anthropic’s Claude via their APIs.
• A reliable bulk upload flow (drag-and-drop or folder select) with progress feedback and graceful error handling.
Deliverables
1. Symfony project with database schema, document ingestor, and embeddings pipeline.
2. Responsive chat UI wired to selected LLM endpoint(s).
3. Clear README covering local setup, environment variables for model keys, and how to add a new model provider.
4. Short test script or Postman collection that shows the PDF ingestion and chat query working end-to-end.
5. Committed to my github repo.
Acceptance criteria
• I can upload 10+ PDFs at once without a timeout.
• Chat replies cite their source pages and stay relevant to the chosen knowledge domain.
• Switching from ChatGPT API to Gemini API or Claude API is a UI toggle, not a rewrite.
• Browser memory stays stable during a 1000-turn conversation.
---
In your bid, post how you would structure the project and any libraries you’d reach for (Symfony Messenger, Mercure, Pinecone, etc.) or I will assume you are a bot.