Edict
Compile and verify agent-oriented programs from JSON AST to WASM with typed effects and formal checks via MCP.
Overview
Edict MCP is an MCP server for the Build phase that lets agents work with a typed, effect-tracked language that compiles JSON AST to WASM with Z3 verification.
What is this MCP server?
- JSON AST input compiled to WASM for portable agent programs
- Typed, effect-tracked language design for safer agent behavior
- Z3-backed verification in the Edict toolchain
- npm package edict-lang 1.9.2 with stdio MCP transport
- GitHub repository Sowiedu/Edict
- Published version 1.9.2
- npm identifier edict-lang with stdio transport
- Pipeline described as JSON AST in, WASM out, with Z3 verification
Community signal: 6 GitHub stars.
What problem does it solve?
Agent logic written only as prompts and loose scripts is hard to test, verify, or deploy consistently as products mature.
Who is it for?
Builders creating custom agent runtimes or verified automation who want WASM output and Z3-backed checks from the agent IDE.
Skip if: Beginners who only need a single REST API wrapper or no-code chat workflows without compilation.
What do I get? / Deliverables
You get a formal agent-language pipeline through MCP so programs can be typed, effect-tracked, and verified before WASM deployment.
- MCP access to Edict compile and verification workflows
- WASM-oriented agent program artifacts from JSON AST
- stdio transport via published npm MCP package
Recommended MCP Servers
Journey fit
Edict is a language and toolchain for building reliable agent logic—work that happens when you are constructing agent runtimes and tooling, not when you are only validating market fit. It targets AI-agent programming (JSON AST in, WASM out) with Z3 verification—squarely agent-tooling in the build catalog shelf.
How it compares
Agent language compiler MCP, not a general code-review or brainstorming skill.
Common Questions / FAQ
Who is Edict MCP for?
Developers shipping agent-heavy products who want a structured language with WASM output and verification instead of unstructured agent scripts.
When should I use Edict MCP?
During Build agent-tooling when you are defining executable agent programs, tracking effects, or validating logic before deployment.
How do I add Edict MCP to my agent?
Install the npm package edict-lang version 1.9.2 and register the stdio MCP server in your Claude Code, Cursor, or Codex MCP configuration.