Workspace Settings

Configure system endpoints, AI orchestration providers, and persistent project memory.

FastAPI Backend Connection

CHECKING

API BASE URL

Points to your FastAPI service hosting LangGraph workflows and database endpoints.

AI Provider Configuration

Google Gemini ProviderStructured extraction & evidence reasoning using Gemini 2.5 Flash

CONFIGURED IN BACKEND

ZERO-FAILURE LOCAL FALLBACK

When GOOGLE_API_KEY is omitted, Nirman automatically uses high-fidelity deterministic reasoning models to prevent blocking local development or offline evaluation.

Database & Vector Storage

Primary store: PostgreSQL / SQLite via SQLAlchemy Async

Supabase schema: supabase/migrations/20250101_initial_schema.sql

Embeddings & similarity: pgvector ready