Application that generates Product Requirements Documents (PDF) from transcript PDFs and creates Linear tickets using MistralAI Agents API with MCP servers.
Use Case
- Extract text from PDF transcripts using OCR
- Generate comprehensive PRDs from meeting transcripts
- Parse PRDs into structured features
- Create Linear tickets automatically for each feature
Architecture
├── mcp_servers/
│ ├── stdio_prd_generator_server.py
│ └── stdio_linear_ticket_gen_server.py
├── app.py
Main Application
- app.py: Chainlit interface with MistralAI agent integration
MCP Servers
- stdio_prd_generator_server.py: PDF OCR processing and PRD generation using MistralAI LLMs
- stdio_linear_ticket_gen_server.py: PRD parsing and Linear ticket creation via GraphQL API
Installation
pip install chainlit mcp loguru pydantic gql
Environment Setup
Set your MistralAI API key:
export MISTRAL_API_KEY="your_api_key_here"
Set your Mistral API key in both the servers:
mcp_servers/stdio_linear_ticket_gen_server.py
mcp_servers/stdio_prd_generator_server
Configure Linear API credentials in mcp_servers/stdio_linear_ticket_gen_server.py
:
- Update
linear_api_key
with your Linear API key - Update
team_id
with your Linear team ID
Usage
Run the application:
chainlit run app.py
Ask questions like:
- "Generate a PRD from transcript.pdf"
- "Generate a PRD outlining LeChat improvements based on the transcript in transcript.pdf, and create corresponding Linear tickets."
The app will automatically extract text from PDFs, generate PRDs, and create corresponding Linear tickets.