Nouli

Nouli is a decision graph CLI and Rust library that runs decision models over multiple turns and feeds past choices into new decisions.

Nouli is a decision graph system that runs decision models over many turns, carrying forward previous facts and decisions to inform future answers. Unlike text-generating LLMs, decision models return probabilities for fixed sets of answers based on provided facts, but operate statelessly. Nouli adds a ledger memory that tracks facts, questions, and responses, allowing sessions to branch and blend together without overwriting past outcomes. It is built for developers and coding agents that need structured, chained decision-making via an interactive prompt, terminal commands, or as an embedded Rust library. Output formats include text, JSON, JSONL, and TOON, with Mermaid graph export options.

Nouli at a glance

Best for: Software engineer
Runs on: CLI, Linux, macOS, open source
Pricing: Free
Released: October 2026
Category: AI coding tool
Website: github.com/standing-desks/nouli

Key features

Branch and blend sessions to combine histories from multiple decision paths
Record sequential facts, questions, and probability-based model answers in a ledger
Run speculative queries without modifying the session ledger using what-if commands
Support choice, score, and yes/no question formats
Export session decision graphs to Mermaid format for visual history tracking
Provide scriptable terminal commands with JSON flags for coding agents
Embed as a native Rust library directly inside other software applications
Output data in text, JSON, JSONL, and TOON formats

Use it for

Orchestrating multi-department recall plans and stock quarantines in food distribution
Determining robotic arm picking, weight checks, and exception actions in warehouse stations
Guiding chained decision trees in lending assessments, credit limits, and seed investing
Managing vulnerability incident responses and energy planning decisions
Enabling autonomous coding agents to feed step-by-step shell findings into structured decisions

Alternatives to Nouli

Decisions API — Fast classification and routing decisions without generating free text.
Enola — Generates a deterministic graph of system architecture locally without AI.
Project STROT — Runs locally at zero marginal cost with typed JSON outputs and confidence-based abstention.
Ollaya — Run decision models locally with typed, calibrated answers in milliseconds.
Laya — It runs typed decisions in a single forward pass over 100+ languages.
Strata — Runs large AI models locally on consumer gaming PCs without server reliance.
Autoloom — It embeds engineering judgment, change necessity checks, and evidence-backed delivery into automated coding.
CodeGraph — It eliminates file crawling, reducing AI agent tool calls by 88% and cutting costs.
Ponytail — Cuts unneeded code and dependencies by enforcing a minimal-code ladder.
MLX Direct Decider — Runs local MLX decisions without fine-tuning or extra weights.
Oracle — Provides a single entry point for multiple pro AI agents and session management.
zerostack — Minimal memory footprint and high CPU efficiency compared to JS-based agents.

All Nouli alternatives