Commit Graph

16 Commits

Author SHA1 Message Date
Adolfo Reyna 84e20f8d81 Apply Hermes optimizations: per-workspace session state file, silent spoken tool activity, clean single-frame turn termination 2026-08-09 21:46:12 -04:00
Adolfo Reyna 2f181ff4f1 Replace OpenCode harness with Hermes, remove workspace dependency, and move vocabulary/corrections to app folder 2026-08-09 21:35:52 -04:00
Adolfo Reyna 757261ebfd Enable live real-time TTS voice synchronization between VoiceManager and KokoroTTS 2026-08-07 20:23:21 -04:00
Adolfo Reyna 5b1aad3fa3 Auto-install bin/voice_tool.py into working workspace directory 2026-08-07 20:18:33 -04:00
Adolfo Reyna 351b87beb8 Add voice management tool bin/voice_tool.py and dynamic voice switching system 2026-08-07 20:16:35 -04:00
Adolfo Reyna 116e2e3763 Enable session persistence with --continue and instruct model to complete tool loops fully 2026-08-07 19:36:36 -04:00
Adolfo Reyna 1386615c49 Enable OpenCode tools with --auto in opencode_llm.py 2026-08-07 19:28:40 -04:00
Adolfo Reyna 7173300298 Add strict truthfulness instruction to VOICE_STYLE to prevent hallucinated tool execution 2026-08-07 19:27:35 -04:00
Adolfo Reyna b510cedbf8 Set default working directory for OpenCode to ~/Workspace 2026-08-07 19:13:34 -04:00
Adolfo Reyna 40a4107bc9 Integrate OpenCode CLI with ollama-cloud/gemma4:31b 2026-08-07 19:10:32 -04:00
Adolfo Reyna 4773b93779 Add OpenCode / Ollama LLM integration with gemma4:31b default 2026-08-07 18:57:11 -04:00
Adolfo Reyna 1a8a93695b Set Qwen2.5-7B-Instruct-4bit as default MLX model and add spoken text filtering 2026-08-07 18:46:05 -04:00
Adolfo Reyna c50da6f29c Set Gemma 2 2B MLX as default local model and add --mlx-model flag 2026-08-07 18:32:57 -04:00
Adolfo Reyna f579a26cc0 Add AGENTS.md template and fix repetition penalty for local macOS LLM 2026-08-07 18:27:51 -04:00
Adolfo Reyna 598d55e383 Add macOS LLM model engine support and environment setup 2026-08-07 18:23:38 -04:00
Adolfo Reyna 62d805e565 Initial commit of current state 2026-08-07 18:15:36 -04:00