difftests 0.1.0

AI-powered test generator from git diffs — write tests for what you changed

difftests

AI test generator from git diffs — write tests for what you just changed.

Stop writing tests from scratch. difftests analyzes your git diff and generates tests that cover exactly what you changed — including edge cases and error conditions.

Install

pip install difftests

Set your API key (or use Ollama for zero-cost local generation:

export ANTHROPIC_API_KEY=your-key   # Claude (default)
export OPENAI_API_KEY=your-key      # or OpenAI
# or use --provider ollama           # local, no API key needed

Usage

# Generate tests for staged changes (most common)
difftests

# Generate tests for all changes vs main branch
difftests --diff main

# Generate tests for a specific file
difftests --file src/auth.py

# Specify the test framework (auto-detected by default)
difftests --framework pytest     # default for Python
difftests --framework jest       # for JavaScript
difftests --framework vitest     # for TypeScript
difftests --framework rspec      # for Ruby
difftests --framework go         # for Go

# Save generated tests to a directory
difftests --output tests/

# Generate AND run tests immediately
difftests --run

# Copy to clipboard
difftests --copy

# Add context for better generation
difftests --context "This is a payment module, be thorough with security tests"

# Use local Ollama (free, no API key)
difftests --provider ollama --model qwen2.5:7b

# Raw output for piping
difftests --raw > tests/test_new.py

Frameworks

Language	Default	Alternatives
Python	pytest	unittest
JavaScript	jest
TypeScript	vitest	jest
Ruby	rspec
Go	go

What Makes difftests Different

Understands your diff, not just your code. difftests focuses on the exact functions and methods you changed, not random parts of your codebase. The AI is shown:

1. The git diff (what changed)
2. The full file content (context)
3. Instructions to focus on edge cases, error conditions, and security

Works with any LLM. Claude, OpenAI, or local Ollama — bring your own model.

Generates real tests, not boilerplate. The AI is explicitly instructed to:

• Test edge cases (null, empty, boundary values)
• Test error conditions (pytest.raises, exception handling)
• Mock external dependencies (databases, HTTP, file I/O)
• Use the framework's best practices (not generic asserts)

Developer Workflow Integration

difftests works best as part of the AI-powered developer workflow:

# 1. Morning: generate standup from yesterday's commits
standup-ai ~/projects/myapp

# 2. Write code, then review before committing
critiq                          # AI review of staged changes
difftests --staged              # generate tests for what you just wrote
pytest tests/                   # run them

# 3. Generate conventional commit message
gpr --commit-run

# 4. Generate PR description
gpr

# 5. Review a teammate's PR (also generate tests for their changes)
difftests --diff main --output tests/
prcat 42

# 6. At release: generate CHANGELOG
changelog-ai --from v0.1.0 --prepend CHANGELOG.md

Providers

Provider	Command	Notes
Claude (default)	--provider claude	Best results; requires ANTHROPIC_API_KEY
OpenAI	--provider openai	Requires OPENAI_API_KEY
Ollama	--provider ollama	Free, runs locally; no API key needed

Related Tools

• critiq — AI code reviewer (find issues before pushing)
• critiq-action — critiq as a GitHub Action for CI
• gpr — AI commit messages + PR descriptions
• prcat — AI reviewer for teammates' pull requests
• gitbrief — git-history-aware context packer for LLMs
• standup-ai — daily standup from git commits
• changelog-ai — AI-generated CHANGELOG
• chronicle — AI git history narrator

License

MIT