Skip to content

JSAT — JaySoft AI Tools

Codebase intelligence shell and SDK. Lightweight by default.

License: MIT Python 3.10+ PyPI

What is JSAT?

JSAT gives any AI (Claude, Codex, GPT, Gemini, Bob Shell, or local Ollama) deep, structured understanding of your codebase — so you spend less time explaining context and more time shipping.

  • Interactive shell — IPython-style REPL for codebase intelligence
  • Python SDK — from jsat import JSAT for programmatic access
  • AI tool integrations — works as an MCP server with Claude Code, Codex, Cursor, Bob Shell, Gemini CLI, and more
  • Works offline — local Ollama or LM Studio, no API keys required
  • Lightweight by default — pip install jsat is ~80 MB and starts in under 800 ms
  • Graph-native — SQLite (solo) or Neo4j (team) codebase graph, indexed by tree-sitter

Quick Start

Three commands to get running:

pip install jsat                  # install
cd your-project && jsat index .   # build the codebase graph
jsat shell                        # open the interactive shell

Inside the shell:

JSAT Shell v0.4.12
> what does this project do?
> which services write to the orders table?
> blast-radius src/payment/refund.py
> switch claude

Choose your AI integration

Native clients, direct Ollama, and clients launched through Ollama are independent routes. Open the tabbed integration chooser for complete setup, model selection, lifecycle, verification, and troubleshooting instructions for each route.

For example, connect native Claude Code with:

jsat connect claude    # wires JSAT as an MCP server + installs /jsat-* commands
jsat claude            # open Claude Code with all JSAT tools active

Then inside Claude Code, use slash commands:

/jsat-query what calls the refund endpoint?
/jsat-blast-radius src/payment/refund.py
/jsat-security
/jsat-incident 500 errors on checkout since 14:00

See Native Claude for the full setup or Claude command reference for every JSAT command surface.

Installation Profiles

Profile Command Size Best for
Core pip install jsat ~80 MB Quick evaluation
Local AI pip install jsat[local] ~85 MB Solo dev with Ollama
Standard pip install jsat[standard] ~120 MB Individual engineers
Team pip install jsat[team] ~200 MB Engineering teams (Neo4j + Qdrant + Redis)
CI pip install jsat[ci] ~90 MB CI pipelines (no AI cost)
Full pip install jsat[all] ~350 MB Power users

Auto-Detection Matrix

JSAT detects your environment on first run and selects the right backends automatically. Run jsat doctor to see what was detected.

Environment Graph Embeddings Cache
Laptop, Ollama running SQLite nomic-embed-code (local) disk
Team server (Neo4j + Qdrant + Redis) Neo4j text-embedding-3-small Redis
Apple M-series / ARM SQLite nomic-embed-code via Metal disk
CI environment SQLite none (skipped) memory
Raspberry Pi (< 4 GB RAM) SQLite nomic-embed-code disk

Supported AI Providers

Provider Needs Cost
Claude Code CLI claude binary installed Free tier available
Anthropic API ANTHROPIC_API_KEY Paid
OpenAI OPENAI_API_KEY Paid
Google Gemini GEMINI_API_KEY Paid
Ollama (local) ollama serve + model Free
LM Studio (local) LM Studio running Free

Next Steps