# fk-peri-code **Repository Path**: g-aicoder/fk-peri-code ## Basic Information - **Project Name**: fk-peri-code - **Description**: 13 MB binary, ~50 MB of RAM, 98% cache hits; 99% AI-generated, docs/superpowers/specs, mattpocock/skills, peri web - **Primary Language**: Rust - **License**: Apache-2.0 - **Default Branch**: main - **Homepage**: https://github.com/KonghaYao/peri - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 0 - **Created**: 2026-07-02 - **Last Updated**: 2026-07-02 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README
# Peri Code **A Rust-built coding agent β€” fast, lean, Claude Code compatible, any LLM.** [![License](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](https://opensource.org/licenses/Apache-2.0) [![Rust](https://img.shields.io/badge/Built%20with-Rust%20πŸ¦€-orange.svg)](https://www.rust-lang.org/) [![Platform](https://img.shields.io/badge/Platform-macOS%20%7C%20Linux%20%7C%20Windows-blue.svg)](#install) [![Models](https://img.shields.io/badge/LLMs-Anthropic%20%7C%20DeepSeek%20%7C%20GLM%20%7C%20Qwen-green.svg)](#why-peri) [![Stars](https://img.shields.io/github/stars/konghayao/peri?style=social)](https://github.com/konghayao/peri)
One **13 MB binary**, **~50 MB of RAM**, **98% cache hits**. Bring your own API key β€” DeepSeek, GLM, Qwen, or Anthropic, switch on the fly. Your existing Claude Code config works today, not eventually: skills, hooks, MCP, plugins, sub-agents. **Zero migration, zero lock-in.** 99% of the codebase was written by AI (DeepSeek & GLM-5.2), shipped by humans who decided *what* to build. The agent files its own bugs, fixes them, and writes the lessons back into the repo. More on that [below](#built-by-ai-published-by-human). --- ## Why Peri - πŸ¦€ **Rust, not Node.js** - 13 MB binary, ~50 MB RAM. Won't sneak up to 1 GB while you're not looking - ⚑ **95–99% cache hit rate** - Frozen system prompt + boundary marker = near-zero wasted tokens - 🌐 **Any LLM** - Anthropic, OpenAI, DeepSeek, GLM, Qwen. Swap mid-session, no restart - πŸͺŸ **Cross-platform** - macOS, Linux, Windows β€” one binary. ConPTY-aware event handling, unified overlay textarea, cross-platform spawn - πŸ”Œ **Drop-in Claude Code compatible** - Existing Claude Code config just works. Skills, hooks, MCP, plugins β€” zero migration - πŸ” **Tool Search** - The LLM only sees what it needs. ~14 core tools, rest discovered on demand β€” lean prompts, fat cache hits - πŸ“ **Streaming Markdown** - Code blocks, tables, diffs β€” fully rendered as the agent types, not after - πŸ€– **Sub-agents & background agents** - 7 built-in specialists (coder, explorer, code-reviewer, web-researcher…). Fork work to the background and keep going - πŸ—œοΈ **Auto Compact** - Hours-long sessions stay fast and cheap, automatically - πŸ“¦ **agm** - `agm install` any skill or agent. One lockfile, any tool - 🌐 **Web Terminal (`peri web`)** - Browser-based remote shell, one command. xterm.js + multi-pane split, auto open browser - πŸ”§ **Built-in LSP & observability** - Language-aware intelligence out of the box. Langfuse traces, token usage, cache monitor --- ## Architecture Peri is not just a TUI. It's a layered platform where the **agent core** is decoupled from the **frontend** via the [Agent Client Protocol](https://agentclientprotocol.com). The same core powers three entry points: ```mermaid graph TD TUI["**peri-tui**
Terminal (ratatui)"] IDE["**Zed / JetBrains**
IDE (ACP client)"] STDIO["**Stdio**
Headless / CI / Cloud"] TUI -->|MpscTransport| ACP IDE -->|ACP Stdio| ACP STDIO -->|ACP Stdio| ACP ACP["**peri-acp** β€” ACP Server
session Β· executor Β· prompt Β· commands"] ACP --> AGENT["**peri-agent**
ReAct loop Β· LLM adapter Β· tools Β· SQLite storage"] ACP --> MW["**peri-middlewares**
18 middlewares: FS Β· HITL Β· SubAgent Β· Skills Β· MCP Β· Hooks Β· Compact"] ACP --> LSP["**peri-lsp**
LSP client"] AGENT -.->|telemetry| LF["**langfuse-client**"] MW -.->|renders with| WIDGETS["**peri-widgets**
Markdown Β· code blocks Β· tables"] ``` **One core, three frontends.** Terminal users get `peri-tui`. IDE users connect via ACP (Zed today, more to come). Headless / CI / cloud scenarios use the Stdio transport. Change the agent logic once β€” every frontend benefits. --- ## Install Binaries available for macOS (x86_64 / Apple Silicon), Linux (x86_64 / aarch64 / riscv64), and Windows (x86_64). ```bash # macOS / Linux curl -fsSL https://raw.githubusercontent.com/konghayao/peri/main/scripts/install.sh | bash # Windows (PowerShell) irm https://raw.githubusercontent.com/konghayao/peri/main/scripts/install.ps1 | iex # start peri peri # self-update peri update ``` First launch guides you through model and API key configuration β€” no config file editing required. --- ## Built by AI, Published by Human Peri's code is 99% AI-generated, primarily by DeepSeek and GLM-5.2. The development workflow is a closed loop the agent drives itself: | When you... | The loop kicks off | |---|---| | **Find a bug or tech debt** | `issue-create` β†’ `systematic-debugging` β†’ `writing-plans` β†’ `subagent-driven-development` β†’ `issue-archive` β†’ update `CLAUDE.md` | | **Want a new feature** | `grill-me` β†’ `writing-plans` β†’ `subagent-driven-development` | | **Codebase getting messy** | `slop-cleaner` β†’ `improve-codebase-architecture` β†’ `writing-plans` β†’ `subagent-driven-development` | Each fix that reveals a non-obvious constraint gets written back into `CLAUDE.md` as a **TRAP** β€” a hard rule the agent follows on every subsequent iteration. The dozens of TRAPs in the repo weren't authored by humans; they were extracted by the agent at the scene of each bug. That's how quality compounds without human code review. β†’ Read the full story: [Nobody Coding](docs/blogs/ai-coding-paradigm/nobody-coding.md) --- --- ## Acknowledgments - [Claude Code Best](https://github.com/claude-code-best/claude-code) β€” community support and feedback - [Superpowers](https://github.com/obra/superpowers) & [Matt Pocock's Skills](https://github.com/mattpocock/skills) β€” the skill suites that drive Peri's AI engineering workflow - [ACP](https://agentclientprotocol.com) β€” open protocol for agent-IDE communication - [rmcp](https://github.com/anthropics/rmcp) β€” Rust MCP client library - [Ratatui](https://ratatui.rs) & [Tokio](https://tokio.rs) - [Langfuse](https://langfuse.com) β€” LLM observability - [Zed](https://zed.dev) β€” first ACP-compatible IDE ## License Apache 2.0