About Iris Code
Iris Code is a local MCP server that gives coding agents project-specific code-quality findings, rules, safe fixes, and gate results. Its tools also cover refactor verification, database schemas, and dependency checks.
Key features
- 01Code-quality findings and measurements
- 02Project rules and thresholds
- 03Preview and apply safe fixes
- 04Refactor planning and verification
- 05Database schema inspection
- 06Quality gates and dependency checks
Use cases
- 01Review a project's code health
- 02Apply mechanical fixes after preview
- 03Check a refactor against a baseline
- 04Inspect dependencies and quality gates
MCP support
The local stdio MCP server exposes Iris Code tools to coding agents, including checks, fixes, refactor workflows, schema inspection, and quality gates.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEIris Code MCP launcher
Give coding agents your project's real code-quality rules, findings, safe fixes and gate results.
This repository contains the npm launcher and setup documentation for Iris Code's local MCP server. It starts the published @iris-code/cli package, which carries the implementation. The analysis engine's source is maintained separately and is not included here. The launcher declares the MIT license; this repository does not claim the whole Iris Code product is open source.
Install
Requires Node.js 18 or newer.
npx -y @iris-code/mcp
Add this to your client's MCP configuration:
{
"mcpServers": {
"iris_code": {
"command": "npx",
"args": ["-y", "@iris-code/mcp"]
}
}
}
Clients normally supply the current workspace root. For a client that does not provide MCP roots, append --root and your absolute project path to args. In JetBrains IDEs, use Tools | Iris Code | Set Up MCP Server from the installed Iris Code plugin.
What agents can do
| Tool | Purpose |
|---|---|
iris_check | Located findings and code-health measurements |
iris_config | Resolved project thresholds and rules |
iris_fix_safe | Preview and apply provable mechanical fixes |
iris_refactor_plan | Located refactor suggestions and a captured baseline |
iris_verify_refactor | Verify improvement against a baseline |
iris_schema | Real database tables, columns and relations |
iris_explain | Rule details behind findings |
iris_gate | Quality-gate result and failing rules |
iris_check_dependency | Dependency advisory evidence |
iris_check_package | Package risk before installation |
Single-file workflows are Free and require no API key. Fresh project work requires sign-in and applicable access. Reopening owned evidence does not spend another project action. Paid access adds broader enforcement and project capabilities. See current pricing and the MCP guide.
Boundaries
Analysis runs locally. The server has no general shell-execution tool. Secret values are masked, and requested paths are confined to the project root, including symlink resolution. Safe fixes preview first; applying needs the preview's single-use token. Dependency and advisory lookups require network consent. Repository handoff policy and access checks still apply.
Support and privacy
Registry metadata release
Version 0.2.2 adds the official MCP Registry ownership name io.github.Daiveedjay/iris-code. Launcher behavior and the CLI dependency range are unchanged. The npm package's repository URL identifies the existing private release owner because npm trusted publishing requires it to match the workflow repository. This public repository supplies launcher source and documentation; npm publication still runs from the private release workflow.
Official MCP Registry identity: io.github.Daiveedjay/iris-code, published version 0.2.2. Verify the public registry metadata. The entry uses local stdio transport, not a hosted endpoint.
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.