dogear

enter for all results · esc to close

Caveman

github.comtool

🦴 A self-contained C# prompt compressor for LLMs (GPT-4, Gemma, Llama, …): strips grammatical "noise" and lemmatizes text for up to 70% token reduction, across 50+ embedded languages, with zero NLP-model runtime dependency. Content-aware routing auto-detects…

from
Blazor
added
2026-10-10
likes
0

similar

  1. Caveman.mcp github.com

    An MCP (Model Context Protocol) server exposing Caveman's token compression, language detection, content-aware routing and extractive summarization as tools for AI agents and MCP-compatible editors — integrates directly with Claude Code, Cursor, Windsurf and similar.

  2. Caveman.SemanticKernel github.com

    Optional add-on package exposing Caveman's prompt compression, conversation-fitting/summarization, and project-analysis tools as Microsoft Semantic Kernel plugins (kernel functions ready to register with Kernel.CreateBuilder()).

  3. Caveman.PrivacyGuard github.com

    🛡️ Enterprise-grade PII/privacy analyzer for AI/LLM pipelines: detects, risk-scores and auto-masks sensitive data across 27 EU countries (heuristic regex + Luhn/IBAN/checksum validators + Shannon-entropy analysis), with GDPR/PCI-DSS/NIST-800-53 compliance mapping and…

  4. Caveman github.com
  5. Fun.Blazor github.com

    This is a project to make F# developer to write blazor easier - Use F# for Blazor, Computation expression (CE) style DSL for internal and third party blazor libraries, Dependency injection, Adaptive and Elmish models, Giraffe style routing, Type safe style.

  6. unslop github.com

    CLI and MCP server that removes AI writing patterns from text: tricolons, em-dash overuse, hedging stacks, sycophancy openers, and overused vocabulary. Works with any MCP client including Gemini CLI. Five intensity levels and a lint-only audit mode. Useful for cleaning commit…

Blazor › Sample Projects > AI: “🦴 A self-contained C# prompt compressor for LLMs (GPT-4, Gemma, Llama, …): strips grammatical "noise" and lemmatizes text for up to 70% token reduction, across 50+ embedded languages, with zero NLP-model runtime dependency. Content-aware routing auto-detects…”