Search and read Empryo's documentation. Read-only, no auth, no local access.
Copy the AI prompt to install this server into Claude Code, Cursor, or another agent β or use 1-click editor setup below.
π‘ Paste the JSON block into your client's configuration file under mcpServers, then restart the application.
previously SoulForge
Code in context.
AI coding with a map of your codebase.
Website Β· Download Β· Benchmarks Β· Changelog Β· Discussions Β· Discord
SoulForge is now Empryo β the same symbol-level, graph-powered agent, rebuilt with a desktop app, a faster engine, and a composable core. This repository is Empryo's public home for issues and discussions.
Official installers and direct downloads are available only from empryo.com/download. Empryo is not distributed through Homebrew, WinGet, or npm.
Desktop app and prebuilt binaries: empryo.com/download. Runs on macOS, Linux, and Windows. Do not download Empryo binaries from GitHub Releases or third-party package managers.
Most coding agents grep, read whole files, and patch strings β they never know what depends on the code they just changed. Empryo builds understanding before it mutates anything:
| Code genome | live dependency graph: tree-sitter across 30+ languages, PageRank + git co-change ranking, blast-radius tags, millisecond search |
| Symbol-level editing | 65+ AST operations (atomic, with rollback) + structural edits in 30+ languages |
| Multi-agent | parallel explore/edit agents with a shared I/O cache β cheap models scout, strong models write |
| Task router | ten routable roles, any model in any seat, per tab β your own mixture of experts |
| Time machine | every prompt is a git checkpoint; rewind code and conversation together, land on any turn |
| Three surfaces | native desktop app, full terminal UI, headless CLI for scripts and CI β one genome, three phenotypes |
| LSP + MCP | 576+ language servers via Mason, any MCP server, 13 lifecycle hooks |
| Free compaction | structural context compaction with no LLM call β long sessions stay cheap |
Empryo isn't one model in a loop β it's a crew, and you assign the seats. Every role is a routable slot that takes any model from any of the 22 providers:
brain Β· spark scout Β· ember code Β· explore Β· verify review Β· goal review Β· desloppify Β· summarize Β· compact Β· web search
Head-to-head against pi β same models, same repositories, same tasks. The graph does more with less:
| Round 1 3 bugs Γ 3 models | Round 2 5 real bugs Β· hono / zod / ky | |
|---|---|---|
| Bugs fixed | 8/9 vs 7/9 | 7/10 vs 6/10 |
| Cost | 28% lower β $1.13 vs $1.58 | 23% lower β $7.08 vs $9.19 |
| Wall-clock | 57% faster β 4m 16s vs 10m | 32% faster β 22m 30s vs 32m 55s |
| Efficiency | 5.7Γ fewer input tokens β 1.09M vs 6.21M | 28% fewer steps β 274 vs 382 |
Round 2 used real bugs from merged PRs (post-training-cutoff, history scrubbed, regression tests injected after each run). Full methodology and transcripts: empryo.com/benchmarks Β· reproduce at proxysoul/pi-vs-empryo-bench.
Empryo runs entirely on your machine. Bring your own key β Anthropic, OpenAI, Google, Groq, DeepSeek, Bedrock, and 16 more, or any OpenAI-compatible endpoint β or run fully local with Ollama / LM Studio. No proxy in the middle, no code leaving your machine, no per-seat fee. Free to use.
SoulForge remains available to download and install, and continues to receive fixes for bugs and critical issues. New features and active development have moved to Empryo.
LICENSE).No reviews yet β be the first to share how this listing worked for you.
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
[](https://allmcps.com/mcp/docs-6)<a href="https://allmcps.com/mcp/docs-6"><img src="https://allmcps.com/api/badge/docs-6?style=directory" alt="Docs on AllMCPs" /></a>