qikly
Generates Python tests from acceptance criteria the coding agent never sees
- Pricing
- Free
- Platforms
- CLI
- Runs
- Locally
- Version
- 0.6.3
- Listing updated
- 3 days ago
About qikly
qikly is a free MCP server for developer tools. It runs locally on your machine and is started by your AI client.
The idea behind qikly is the exam analogy: if the same model writes the code and the checks, everything goes green and the green means little. qikly splits the job in two. A test agent builds the suite from acceptance criteria, while a separate coding agent works toward passing it without ever reading those criteria, and an automated repair loop iterates until the tests pass. In the logged example, the coding agent discovered a rate limit of 100 from a failing test and added the check itself.
The workflow has three steps, each taking a module or a whole folder. Scoring plants faults in a copy of your code and reports how many your existing tests miss, turning each miss into a question about unchecked behaviour. Scaffolding drafts a task file of acceptance criteria from your README and docstrings, never from the code, and marks guesses for review. The final step runs the loop.
Scoring and validation need no API key; the later steps use a Gemini, OpenAI or Anthropic key. A --demo mode runs the whole loop on a bundled task in about thirty seconds. The README reports a run over 50 young Python libraries that found four real problems. It targets functional tests for backend Python.
Install
Run in a terminal
claude mcp add qikly -- uvx qiklyAdd to claude_desktop_config.json
{
"mcpServers": {
"qikly": {
"command": "uvx",
"args": [
"qikly"
]
}
}
}Add to ~/.cursor/mcp.json
{
"mcpServers": {
"qikly": {
"command": "uvx",
"args": [
"qikly"
]
}
}
}Add to .vscode/mcp.json
{
"servers": {
"qikly": {
"type": "stdio",
"command": "uvx",
"args": [
"qikly"
]
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"qikly": {
"command": "uvx",
"args": [
"qikly"
]
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"qikly": {
"command": "uvx",
"args": [
"qikly"
]
}
}
}Configuration
QIKLY_PROJECT_ROOTThe folder holding your tasks, the one with inputs_private/. Set it when your editor has a different folder open, which is common.
What it does
- Independent test generation
- A test agent writes tests from criteria the coding agent cannot see.
- Fault-planting score
- Measures how many planted faults your current tests miss.
- Spec drafting
- Drafts acceptance criteria from README and docstrings, not code.
- Repair loop
- Iterates the coding agent until the tests pass.
- Demo mode
- qikly --demo runs the full loop on a bundled task.
- CI integration
- The README lists a GitHub Action for test generation.
Good for
- Check whether AI-written tests really constrain behaviour
- Find untested behaviour in a Python module
- Draft acceptance criteria from existing docs
- Let a coding agent converge on code against hidden tests
Pricing
Free
Free to install; model calls use your own API key
Prices change. Check test.qikly.com before you buy.
qikly FAQ
Do I need an API key?
Not for scoring or validating. Scaffolding and the full loop need a Gemini, OpenAI or Anthropic key; Gemini's free tier needs no card.
What language does it cover?
Functional tests for backend Python.
Does scoring change my code?
No, faults are planted in a copy.
How much does the demo cost?
The site says about thirty seconds and well under a cent with an API key.
Is qikly free?
Yes. qikly is free to use.
How do I add qikly to Claude Code?
Run this command in your terminal: claude mcp add qikly -- uvx qikly
What are the best alternatives to qikly?
The closest alternatives to qikly are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.
Alternatives to qikly

REA
MCP serverCLI and MCP server that lets coding agents reverse engineer apps and binaries
Runs locally- Open source
- 464K GitHub stars
- Developer tools

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

GitHub
MCP serverOfficial MCP server connecting AI assistants to GitHub repos, issues and workflows
Not checked yet- Remote
- Open source
- 34.5K GitHub stars
- Developer tools

Serena MCP: the IDE for your agent
MCP serverMCP toolkit giving coding agents symbol-level code retrieval and editing
Runs locally- Open source
- 31.5K GitHub stars
- Developer tools

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Ouroboros
MCP serverSpec-first layer that pins requirements and verifies results of AI coding agent work
Runs locally- Open source
- 6.4K GitHub stars
- Developer tools