Model Used

The application uses OpenRouter with:

openrouter/free- openai/gpt-oss-20b:free

The model can be changed using the OPENROUTER_MODEL environment variable.

NimitAI Transcript Signals

A small web app that analyzes a meeting transcript and returns AI-detected sales signals with a coaching tip for each one.

What it does

• POST /analyse accepts a JSON payload with transcript
• The backend calls OpenRouter and asks for clean JSON only
• The frontend lets you paste a transcript and shows results as simple cards

How it works

sequenceDiagram
    participant User
    participant Frontend (JS)
    participant Backend (Python)
    participant OpenRouter (AI)

    User->>Frontend (JS): Paste transcript & Click "Analyse"
    Frontend (JS)->>Backend (Python): POST /analyse { transcript }
    alt API Key Present
        Backend (Python)->>OpenRouter (AI): Request Analysis (LLM)
        OpenRouter (AI)-->>Backend (Python): JSON Signals + Tips
    else No API Key
        Backend (Python)->>Backend (Python): Return empty signals
    end
    Backend (Python)-->>Frontend (JS): JSON Response
    Frontend (JS)->>User: Render Signal Cards (Buying intent, Objections, etc.)

Loading

Setup

1. Set OPENROUTER_API_KEY in your environment for live LLM analysis.
2. Run python main.py.
3. Open http://localhost:8000.

Environment variables

• OPENROUTER_API_KEY is used for live LLM analysis; if it is missing, the app returns no signals
• OPENROUTER_MODEL is optional and defaults to openrouter/free
• PORT is optional and defaults to 8000
• OPENROUTER_SITE_URL and OPENROUTER_APP_NAME are optional metadata headers for OpenRouter

LLM used

OpenRouter, via the OpenAI-compatible Chat Completions API.

Notes

If OPENROUTER_API_KEY or OPENAI_API_KEY is not set, the app returns an empty signal list.