Pinned Loading

1. lovework-public lovework-public Public

   Personal AI agent for continuous career discovery. Sources opportunities from email, LinkedIn, and job boards; scores them against your skills and preferences; learns from your decisions to find be…

   Python

2. hermes-advisor-plugin hermes-advisor-plugin Public

   A persistent second-model review plugin for Hermes Agent. Port of pi-omplike-advisor. IQ tokens too cheap to meter with fast and capable mid-models DeepSeek-V4-Flash and Mimo-v2.5, with 1M contexts…

   Python 4

3. ChEMBLdb-query ChEMBLdb-query Public

   LLM-assisted SQL querying of ChEMBL with a judge-loop for self-improvement, synthetic data generation, and bidirectional query↔semantic-plan translation.

   Python 3

4. ds4 ds4 Public

   Forked from antirez/ds4

   DeepSeek-V4-Flash local inference engine for Apple Silicon (Metal). A fork that makes it possible to run the REAP25 version on a 3-year-old M2 Max with 96GB — production-grade DeepSeek-V4-Flash inf…

   C 6

5. Telegram-voice-text-chat Telegram-voice-text-chat Public

   Multilingual automated customer support bot for Telegram with LLM intent/slot extraction. English, French, Wolof.

   Python 1

6. testing123-asi-rocm-local-llamas testing123-asi-rocm-local-llamas Public

   Local LLMs on local hardware I own. Ongoing testing and benchmarking of new models and hardware as they come along. Always an incumbent best model / server / harness for whatever hardware is on hand.

   Shell 2