FP

fmxexpress/pasclaw

Developer tools

이 페이지는 AI로 자동 번역되었습니다. 기술적인 세부 사항은 원문을 확인하세요. 원문 보기

64 stars 품질 80 트렌드 34MIT

Ultra-lightweight native AI agent framework written entirely in Object Pascal for Delphi and Free Pascal, designed for performance, portability, and extensibility.

개요

PasClaw is an ultra-lightweight personal AI agent written in Delphi Object Pascal. It is a Delphi/FPC port inspired by picoclaw, with a command-line assistant, tool calling, MCP integration, an HTTP gateway, an OpenAI-compatible API surface, a small embedded web UI, scheduled tasks, skills, and channel integrations. The main program lives at src/pasclaw/PasClaw.dpr. It initializes terminal color handling, prints the banner, applies timezone configuration, and dispatches into the command tree implemented under src/cmd/. — the fastest way to try PasClaw without building anything locally: The image's entrypoint runs pasclaw for you, so pass only the subcommand (agent, serve, …) — or nothing, which starts the gateway. The gateway's web UI has a first-boot onboarding wizard, so you pick a provider and paste a key in the browser — nothing on the command line. To keep config + workspace across runs, add -v "$HOME/.pasclaw:/home/pasclaw/.

README

원문 README의 일부입니다. 원문 보기.

PasClaw

PasClaw is an ultra-lightweight personal AI agent written in Delphi Object Pascal. It is a Delphi/FPC port inspired by picoclaw, with a command-line assistant, tool calling, MCP integration, an HTTP gateway, an OpenAI-compatible API surface, a small embedded web UI, scheduled tasks, skills, and channel integrations.

The main program lives at src/pasclaw/PasClaw.dpr. It initializes terminal color handling, prints the banner, applies timezone configuration, and dispatches into the command tree implemented under src/cmd/.

Quick start

Docker — the fastest way to try PasClaw without building anything locally:

docker pull fmxexpress/pasclaw

# HTTP gateway + web UI on http://localhost:8088 -- onboard right in the browser.
# The image's default command already runs the gateway, so no arguments needed.
docker run -p 8088:8088 fmxexpress/pasclaw

# Interactive CLI agent (run `pasclaw onboard` inside, or set a provider key)
docker run -it fmxexpress/pasclaw agent

The image’s entrypoint runs pasclaw for you, so pass only the subcommand (agent, serve, …) — or nothing, which starts the gateway. The gateway’s web UI has a first-boot onboarding wizard, so you pick a provider and paste a key in the browser — nothing on the command line. To keep config + workspace across runs, add -v "$HOME/.pasclaw:/home/pasclaw/.pasclaw" (the container’s PASCLAW_HOME, owned by its pasclaw user). See docker/README.md for more. (This runs PasClaw itself in a container — distinct from the optional Docker shell-backend, which runs the model’s shell_exec calls inside a throwaway container for a stronger sandbox.)

Native binary — build the single ~4–5 MB binary (no runtime deps beyond libsqlite3 and OpenSSL):

make get-indy && make          # FPC  -> build/pasclaw
pasclaw onboard                # pick a provider, paste a key, choose options
pasclaw agent                  # start chatting

See Build for Delphi / cross-compile / Windows-on-ARM.

Features

LLM providers — ~26-entry catalog covering Anthropic Messages, OpenAI Chat Completions, Google Gemini generateContent, plus OpenAI-protocol-compatible providers (Groq, DeepSeek, Mistral, Cerebras, OpenRouter, Zhipu, Qwen, Moonshot, MiniMax, NVIDIA NIM, Novita, MiMo, VolcEngine, xAI/Grok, Perplexity, Ollama, LM Studio, vLLM, LiteLLM) and three Cloudflare AI Gateway fronts (cloudflare, cloudflare-anthropic, cloudflare-gemini). The full list lives in src/pkg/providers/PasClaw.Providers.Catalog.pas; adding a provider is a one-record append. Streaming SSE on all three protocol families; native search grounding on Gemini; citation collation on Perplexity. Code-driven SetProvider('anthropic', $ANTHROPIC_API_KEY) for embedders so no ~/.pasclaw/config.json is required.

GitHub에서 전체 README 보기

추천 도구

다른 키워드를 입력하거나 필터를 제거해 보세요.

설치

npx skillfish add fmxexpress/pasclaw
MCP 초보자라면? mcpskill.org 가이드부터 시작하세요 →