Providers & Models
TomoriBot doesn’t have a built-in AI brain — you connect one. A provider is an AI service (Google Gemini, OpenRouter, NovelAI, a local endpoint, …), and a model is a specific model on that provider. You need at least one provider to use her at all.
API Keys
Section titled “API Keys”Add a provider key during first-time setup with /config setup, or later with
/config api-key set. Keys are encrypted at rest — no one, including server admins, can
read them back.
Each provider has its own key-generation steps. Run /help api-key and pick your
provider for the exact walkthrough, or use these starting points:
| Provider | Notes | Get a key |
|---|---|---|
| Google Gemini | Free tier, runs every feature. Recommended first setup. | AI Studio |
| OpenRouter | One key, many models (some free). | OpenRouter keys |
| NovelAI | Subscription; uncensored storytelling/roleplay (text only). | NovelAI |
| DeepSeek | Pay-as-you-go reasoning models. | DeepSeek |
| NVIDIA NIM | Hosted text, embeddings, and image. | NVIDIA Build |
| Anthropic | Claude models via the API (not Claude Code). | — |
| Z.ai | GLM family. ⚠️ ToS restricts usage to coding/agent scenarios. | Z.ai |
| Vertex AI | Google Cloud via gcloud ADC — best for locally-run/dev setups. |
see below |
| Vertex AI Express | Google Cloud API-key BYOK (Preview, Gemini subset). | Express Mode |
| Custom | Any OpenAI-compatible endpoint (Ollama, vLLM, LiteLLM, …). | see Custom Endpoints |
Vertex AI authenticates with Application Default Credentials rather than a stored secret.
For local hosting, ADC can come from gcloud; hosted deployments should use a workload identity
or service account. An AI Studio API key alone does not authenticate full Vertex AI. The selected
project must have billing and the Vertex AI API enabled, and the host identity needs Vertex access.
The full setup guide lives in /help api-key provider:Google Vertex AI.
Google-backed provider setup validates credentials through the authenticated model-listing endpoint. It does not generate text or depend on whichever chat model is currently marked as the catalog default, so a retired default cannot prevent a valid credential from being saved.
Optional: Brave Search key
Section titled “Optional: Brave Search key”Brave Search is separate from your AI provider and only enhances web search (adds image,
video, and news search). Set it with /optional-key brave set. ⚠️ Brave includes $5/month
free credit — set a $5 usage limit in the Brave dashboard to avoid charges.
Choosing Models
Section titled “Choosing Models”/provider and /model are server-scoped: they set the shared defaults every member of
this server uses, and they need the required server permission. Individual members can override
those defaults for their own requests with /personal provider, which follows them across every
server they use TomoriBot in. See
Personalization for that side.
After a provider is set, pick which model each capability uses for this server:
/model text— the main chat model/model vision— a vision model (for reading images when the chat model can’t)/model image— image generation (see Image Generation)/model video— video generation/model embedding— embeddings for the document knowledge base/model speech//model transcription— voice
You can also manage this server’s backup keys for automatic failover and load balancing with
/provider api-key rotation.
Custom Endpoints
Section titled “Custom Endpoints”Custom endpoints let you register self-hosted or proxy-backed services — Ollama, LM Studio, LiteLLM, vLLM, ComfyUI, local TTS/STT — as labeled provider bundles.
- Server scope:
/provider custom-endpoint add/remove. - Personal scope:
/personal custom-endpoint add/remove(just you — see Personalization).
A label groups every capability under one bundle. After registering, select the label
from /model text, /model image, /model video, etc.; if a label has several models for a
capability, a picker lets you choose. Re-run the add command with the same label and
capability but a different model name to register an additional model on that connection (its
URL and API style are inherited).
For full walkthroughs of running the servers, see:
- Setup: Local LLM — Ollama, KoboldCPP, LM Studio, vLLM, LiteLLM.
- Setup: ComfyUI — local image/video generation.
- Setup: ChatMock — ChatGPT account / Codex CLI.
Supported Providers
Section titled “Supported Providers”If you don’t have the hardware to host your own models, TomoriBot supports a wide range of services. Not every feature is available on every provider.
LLM Providers
Section titled “LLM Providers”| Provider | Streaming | Tool Calling | Image Input | Embeddings | Notes |
|---|---|---|---|---|---|
| Google Gemini | ✅ | ✅ | ✅ | ✅ | Free models available |
| OpenRouter | ✅ | ✅ | ✅ | ✅ | Free models available |
| Anthropic (API) | ✅ | ✅ | ✅ | – | Not Claude Code |
| NovelAI | ✅ | ✅ | – | – | Only GLM 4.6 can use tools |
| NVIDIA NIM | ✅ | ✅ | ✅ | ✅ | Free models available |
| DeepSeek | ✅ | ✅ | – | – | – |
| Z.ai | ✅ | ✅ | ✅ | – | Free models; ⚠️ ToS = coding/agent use only |
| Z.ai Coding | ✅ | ✅ | – | – | Subscription plan |
| Google Vertex AI | ✅ | ✅ | ✅ | ✅ | Includes ‘free’ Express version |
| Codex CLI (via ChatMock) | ✅ | ✅ | ✅ | – | Setup |
Image Generation
Section titled “Image Generation”| Provider | Text-to-Image | Image-to-Image | Inpainting | Notes |
|---|---|---|---|---|
| ✅ | ✅ | – | – | |
| OpenRouter | ✅ | ✅ | – | – |
| NovelAI | ✅ | ✅ | ✅ | Can combine with other providers |
| NVIDIA | ✅ | – | – | Text-to-image only; reference images are ignored |
| Z.ai | ✅ | – | – | – |
Video Generation
Section titled “Video Generation”| Provider | Text-to-Video | Image-to-Video | Notes |
|---|---|---|---|
| ✅ | ✅ | Async polling workflow | |
| OpenRouter | ✅ | ✅ | Async polling workflow |
| Z.ai | ✅ | ✅ | Async polling workflow |
Voice & Audio
Section titled “Voice & Audio”| Provider | Text-to-Speech | Speech-to-Text |
|---|---|---|
| ElevenLabs | ✅ | ✅ |
Local voice engines are covered under Self-Hosting. For the built-in web search and URL-fetch engines, see Tools & Extensions.