- Category
- AI Tools
- Rank
- No. 1276Tools index
Previous survey · No. 1283 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- ggml-org
- GitHub
- 194 stars
- Date
About
LLM-based code completion engine from the ggml team — local-first autocomplete for editors using small models.
What it does
P1 serves mainly as a project hub. Its documentation points readers to separate Vim and VS Code implementations, without explaining installation, configuration, model handling, or completion behavior.
Why it's ranked here
The idea has concrete editor targets, but this repository provides too little substance for confident adoption. Its value depends almost entirely on following links to separate implementation projects.
What's good
The project names two practical integration routes: Vim and VS Code. That gives interested developers clear places to continue investigating instead of presenting only a concept.
Tradeoffs
The available documentation is extremely sparse. It offers no setup guide, supported platforms, architecture, model requirements, usage examples, license details, or operational guarantees.
How to use it well
Treat P1 as a discovery point for editor-specific implementations. It suits developers exploring those two editors, but it does not itself document a complete installation or integration workflow.
Technical notes+
README.md contains a heading describing p1 as an edge code-completion engine and links to two external implementation repositories, ggml-org/llama.vim and ggml-org/llama.vscode. The supplied repository text includes no source code, package manifest, configuration, tests, API description, installation commands, or license declaration.
Observed
- Repository surface
- The supplied repository text contains only README documentation.
- Editor interfaces
- The documentation identifies Vim and VS Code implementations.
- Implementation structure
- Both implementations are linked as separate ggml-org repositories.
- Installation surface
- No installation instructions or package manager commands are provided.
- Testing structure
- No tests or test directory appear in the supplied repository text.
Read from README.md.
What it can do
Provide code autocompletion suggestions
Partial code or cursor position in code editor → Code completion suggestions
Generate code locally without internet connection
Code context and partial input → Generated code snippets
Integrate with code editors for real-time assistance
Editor integration and active code file → Inline code suggestions
Process code completion requests using small language models
Code context and completion request → AI-generated code completions
Run inference on local machine without cloud dependency
Code editing session → Local code completion results
Tags
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.