Show and tell: share your qq-ai-bot deployment, stack, or screenshots

[happysnaker]

If you deployed qq-ai-bot for a real chat flow, feel free to share it here.

High-signal things to include:

• which QQ / OneBot side you used (NapCat, LLOneBot, etc.)
• whether the AI side is a mock agent, traecli, or your own ACP-compatible runtime
• whether you put DeepSeek, Ollama, or another local / self-hosted model behind that runtime
• anything you changed for session storage, prompts, observability, or deployment shape
• one screenshot / log snippet / compose fragment if you want

A compact template is enough:

OneBot side:
AI runtime:
Model provider (optional):
Interesting deployment detail:
Screenshot / repo / compose link (optional):

I may link the first few concrete community examples from the landing page so new users can see real deployment shapes instead of only README copy.

[happysnaker]

Maintainer seed example:

• OneBot side: NapCat
• AI runtime: repo-built mock ACP agent
• Model provider: none (this is the deterministic first-hop validation path)
• Interesting deployment detail: the documented Docker demo keeps the whole chain inside one reproducible stack — NapCat -> qq-ai-bot -> mock ACP agent — so people can validate transport, session persistence, and progress streaming before they swap in a real agent runtime.
• Useful entry points: docs/docker-quickstart.md, docs/deployment-patterns.md, and the public project page

If anyone ships a real DeepSeek / Ollama / custom ACP runtime behind this, that is exactly the kind of screenshot / compose fragment / note I want collected here.

[happysnaker]

If you want to drop a quick public deployment note here, the easiest useful format is:\n\n- OneBot side: NapCat / LLOneBot\n- WS mode: reverse / forward\n- Runtime shape: Docker / local / other\n- Agent side: mock ACP / traecli / custom ACP agent / other\n- Model side: none / DeepSeek / Ollama / other\n- One thing that bit you once: token / WS path / container networking / media handling / session boundary / other\n- One thing that felt surprisingly smooth: ...\n\nEven a 5-line report is useful. I am using this thread as public evidence for which deployment shapes are actually working outside my own machine, and I will fold good reports back into the docs / deployment matrix.

[happysnaker]

Small new operator path: a CasaOS app-store PR is now open for qq-ai-bot, bundling NapCat WebUI + the bridge service together so the install path is more browser-first than raw compose copy/paste.

If you try it through CasaOS / a homelab dashboard / a self-hosted app directory, that is exactly the kind of install report I want here:

• what host / NAS / mini-PC you used
• whether NapCat login + QQ sign-in was the only real friction point
• whether the bundled mock ACP agent was enough to validate the stack before swapping in your own agent
• screenshots are welcome, especially if you came from a homelab / CasaOS context

Open PR: Cp0204/CasaOS-AppStore-Play#42