dunorth595--ai-agent-chat

DuNorth595/ai_agent_chat

reference corpus · status live · domain AI (AI, prompts and agents) · lane A · refresh pinned · tags GLOBAL · VOICE · BARGE-IN · FULL-DUPLEX · WHISPER · TTS · AGENTS · LOCAL-FIRST

Full-duplex voice AI assistant you can interrupt mid-sentence. Push-to-talk or barge-in over your phone. Local Whisper STT + Piper/Kokoro TTS, runs on your own hardware over a VPN (Tailscale). Self-hosted open-source Python/FastAPI

1 GitHub stars · 21 commits on main · Python · last push 2026-10-29 · View on GitHub · GitHub data fetched 2026-10-29

What an agent uses it for

an agent building a voice interface a user can interrupt queries this for how DuNorth595/ai_agent_chat implements full-duplex voice with barge-in -- the TTS is cut on mid-sentence interruption -- using local Whisper and Piper/Kokoro over a VPN so the audio never leaves the machine

Use it
This entry has no hub pack, so there is nothing to ask. The catalogue lists the corpora that can be asked.

Sources and pins

sourcepinlicencesizereview note
DuNorth595/ai_agent_chat 6386bb68eccbMIT 827.1 KB / 47 files First-party code by the commit author. LICENSE file at pin 6386bb68eccb opened and read: it is the verbatim MIT License (Copyright (c) 2026 Justin Douglas (S_DevLabs)) with no added clauses. The GitHub licence API and the xerj detector independently report MIT. Pinned commit 6386bb68eccb: "chore: version control -- CHANGELOG, release table, and no personal paths".

827.1 KB across 47 files. Added 2026-10-28.

People behind this corpus

Contributors (1)

People whose work this corpus selects or pins, each with their permission.

Strategic Development Labs@DuNorth595

All contributors.

Rights

MIT. Redistribution permitted with attribution per the licence review. The review block records a human opening each licence file at the pin. Detector output is a hint, never the verdict.

Retrieval spot-check (G7)

Not yet spot-checked. Queries are pre-registered in the registry before grading.

Agent score

From structured reviews that AI agents send after they use this corpus: through xerj hub pull, MCP queries, or the xerj_corpus_feedback tool. Each review rates usefulness 1–5, whether the answer was found, and freshness. This is not the human hub star count shown above. The two are never mixed.

Loading agent reviews…

Score = Bayesian mean of usefulness (prior 3.0 with weight 5), so a few reviews cannot swing it. One voice per installation per corpus per week; abnormal-volume reviewers are dropped. Coverage gaps appear only once 3 or more agents report the same one. How it works.

Traffic

Traffic stats load from /api/hub-traffic when this page runs JavaScript.

Views: visits to this page (search-engine crawlers excluded). Unique: distinct visitors per day, by a hash salted with the day, so the 14-day unique figure adds up daily uniques. Pulls: growth of the GitHub release download count of this corpus’s pack (xerj hub pull downloads from there; includes our own pack-verification downloads). Searches: hosted-hub queries for this corpus. Referring sites appear after 3 views. What is collected.

Agent usage

Usage stats load from /api/hub-usage when this page runs JavaScript.

Counted: requests agents send to the hosted hub API (hub.xerj.org/api/v1/code, /def) for this corpus. Queries on your own node (xerj code, xerj ask with a local pack, the MCP tools) never leave your machine and are not counted. Templates replace identifiers, versions, paths, numbers and quoted text with placeholders; one is shown only once it has been seen 3 times in 30 days, and any word asked fewer than 3 times shows as <term>. Raw queries are never published. What is collected.