NeuriCo#

Autonomous AI research framework. Idea in, paper out.

Quick Reference#

What it doesTakes a research idea (YAML) and autonomously runs the full research lifecycle: literature review, experiment design, code execution, analysis, paper writing, GitHub pushInputYAML file with 3 required fields: title, domain, hypothesisOutputCode ( src/), results & plots ( results/), LaTeX paper ( paper_draft/), GitHub repoProvidersClaude Code, Codex, Gemini (OAuth login, not API keys)Installgit clone https://github.com/ChicagoHAI/neurico && cd neurico && ./neurico setupSourcegithub.com/ChicagoHAI/neurico — Chicago Human+AI Lab (ChicagoHAI), University of ChicagoLicenseApache 2.0

Requirements#

Minimal (one of)#

OptionWhat you needDocker (recommended)git + dockerNativegit + python>=3.10 + uv

Resource#

Access to at least one AI coding CLI (OAuth login required):

• Claude Code (recommended)
• Codex
• Gemini CLI

Recommended#

WhatWhyGitHub token (classic, repo scope)Auto-creates repos and pushes results. Create here

Optional API Keys#

KeyPurposeOPENAI_API_KEYLLM-based repo naming, IdeaHub fetching, paper-finderS2_API_KEYSemantic Scholar literature search via paper-finderOPENROUTER_KEYMulti-model access during experimentsCOHERE_API_KEYImproves paper-finder ranking (~7% boost)HF_TOKENHugging Face private models/datasetsWANDB_API_KEYWeights & Biases experiment tracking

Setup Tiers#

• Basic: CLI login + GITHUB_TOKEN -- full NeuriCo functionality
• Enhanced: + OPENAI_API_KEY -- LLM repo naming + IdeaHub support
• Full: + S2_API_KEY (+ optional COHERE_API_KEY) -- paper-finder literature search

Installation#

Docker (recommended)#

The Docker image is a pre-configured environment with Python, Node.js, AI coding CLIs (Claude Code, Codex, Gemini), and a full LaTeX installation for paper compilation -- so you don't have to install any of these yourself. All experiments run inside this container; nothing is installed on your host system beyond the cloned repo. The image is built from the open-source Dockerfile and hosted on GitHub Container Registry.

bash

git clone https://github.com/ChicagoHAI/neurico && cd neurico
./neurico setup     # pulls Docker image, configures API keys, walks through CLI login

Or step by step:

bash

git clone https://github.com/ChicagoHAI/neurico && cd neurico
docker pull ghcr.io/chicagohai/neurico:latest
docker tag ghcr.io/chicagohai/neurico:latest chicagohai/neurico:latest
./neurico config    # configure API keys
claude              # login to AI CLI (one-time, on host)

Native#

bash

git clone https://github.com/ChicagoHAI/neurico && cd neurico
curl -LsSf https://astral.sh/uv/install.sh | sh
uv sync
cp .env.example .env   # edit: add your API keys
claude                  # login to AI CLI

Invocation#

Fastest: Fetch from IdeaHub and run#

bash

./neurico fetch <ideahub_url> --submit --run --provider claude

Browse ideas at IdeaHub, copy the URL, and run the command above. NeuriCo fetches the idea, creates a GitHub repo, runs experiments, writes a paper, and pushes everything.

From a YAML file#

bash

./neurico submit path/to/idea.yaml
./neurico run <idea_id> --provider claude

Run options#

OptionDescription--provider claude|gemini|codexAI provider (default: claude)--no-githubRun locally without GitHub integration--write-paperGenerate LaTeX paper after experiments (default: on)--paper-style neurips|icml|acl|amsPaper format (default: neurips)--privateCreate private GitHub repository

Input Format#

Only 3 fields required:

yaml

idea:
  title: "Do LLMs understand causality?"
  domain: artificial_intelligence
  hypothesis: "LLMs can distinguish causal from correlational relationships"

Optional fields: background (papers, datasets, code references), methodology (approach, steps, baselines, metrics), constraints (compute, time, memory, budget), expected_outputs, evaluation_criteria.

Full schema: ideas/schema.yaml

Output Format#

text

workspace/<repo-name>/
  src/            # Python experiment code
  results/        # Metrics, plots, models
  paper_draft/    # LaTeX paper (with --write-paper)
  logs/           # Execution logs
  artifacts/      # Models, checkpoints
  .neurico/       # Original idea spec

Results are automatically pushed to the GitHub repo created during submission.

Supported Domains#

DomainExamplesArtificial IntelligenceLLM evaluation, prompt engineering, AI agentsMachine LearningTraining, evaluation, hyperparameter tuningData ScienceEDA, statistical analysis, visualizationNLPLanguage model experiments, text analysisComputer VisionImage processing, object detectionReinforcement LearningAgent training, policy evaluationSystemsPerformance benchmarking, optimizationTheoryAlgorithmic analysis, proof verificationScientific ComputingSimulations, numerical methods

Configuration#

bash

./neurico config      # Interactive API key configuration
./neurico setup       # Full setup wizard
./neurico shell       # Interactive shell inside container
./neurico help        # Show all commands

Environment variables go in .env (copy from .env.example). See README for details.

Security#

• No secrets are uploaded. API keys and tokens stay local in your .env file and are never committed, pushed, or sent anywhere beyond the APIs they authenticate with. Sensitive environment variables are explicitly filtered out from all subprocess calls and sanitized from logs.
• Experiments run inside Docker. The container is isolated from your host system. The only host directories mounted are your config, templates, and workspace output folder.
• Open source. The entire codebase, including the Dockerfile and install script, is publicly auditable on GitHub.
• Built by ChicagoHAI — the Human+AI Lab at the University of Chicago.