DaoXESillyTavern connects to gateways through Chat Completion → Custom (OpenAI-compatible): a Custom Endpoint (Base URL), a key, and a model ID. The UI itself warns about the two classic failure modes — a missing /v1 and a hand-added /chat/completions suffix — so read the setup below before typing anything (cheapest APIs guide).
Why route SillyTavern through DaoXE#
- Built for deep customization. Character cards, lorebooks, instruct templates and group chats all run on one endpoint.
- One key, many vendors. Switch model IDs between chats without touching the endpoint.
- Server-side proxy. The Node server relays requests, so CORS never blocks a browser call.
Set it up in SillyTavern#
- Open the API Connections panel. Top Plug icon → API Connections. Set the API to Chat Completion.
- Choose the Custom source. Set Chat Completion Source to Custom (OpenAI-compatible).
- Fill endpoint and key. Custom Endpoint (Base URL):
https://daoxe.com/v1— do NOT add/chat/completions, ST appends it. Paste your DaoXE key under Custom API Key. - Pick the model, connect, test. Type an exact ID from
GET /v1/modelsin Enter a Model ID (or pick from the dropdown if it populates), click Connect, then run the Test Message check.
Caveats worth knowing
The /v1 rule cuts both ways: ST appends /chat/completions to what you type — the UI hint literally says "Doesn't work? Try adding /v1 at the end!" and warns the suffix is added automatically. A bare https://daoxe.com gives 404s; the correct value ends at /v1. Prompt post-processing: if a chat errors with role/sequence complaints (some OpenAI-shaped servers only accept strictly alternating messages), set Custom Prompt Post-Processing — the Strict or Semi-strict variants rewrite the history accordingly. Tool calling: the post-processing variants marked "no tools" strip tool calls — use a with-tools variant if your model does function calling. Bypass API status check exists for endpoints that work but flunk ST's status probe — try it before rewriting the URL. Exact IDs from GET /v1/models; guessed names fail at first message. More clients: client setup notes.
Verify you actually get the model#
Prove the endpoint works before blaming the client — if this fails, no setting will fix it:
export DAOXE_API_KEY="your_api_key"
# List the exact model IDs your account can call
curl --fail-with-body --show-error --silent \
https://daoxe.com/v1/models \
-H "Authorization: Bearer ${DAOXE_API_KEY}"Prove the endpoint outside ST first, then diff a hard prompt against the official API at temperature 0:
curl --fail-with-body --show-error --silent \
https://daoxe.com/v1/chat/completions \
-H "Authorization: Bearer ${DAOXE_API_KEY}" \
-H "Content-Type: application/json" \
-d '{
"model": "YOUR_EXACT_MODEL_ID",
"max_tokens": 64,
"messages": [{"role": "user", "content": "Say hello in one sentence."}]
}'Verify us — don't trust us
Point the open benchmark at DaoXE and at the official API and compare at temperature 0. Then learn to detect model swapping so a cheaper endpoint can't quietly swap you to a smaller model.
Frequently asked questions#
Can SillyTavern use DaoXE?
Yes — Chat Completion → Custom (OpenAI-compatible): Base URL https://daoxe.com/v1, your key, an exact model ID.
Do I add /chat/completions to the URL?
No — SillyTavern appends it. The Base URL ends at /v1; adding the suffix by hand is the most common 404.
The model list won't populate — what then?
Type the model ID manually in Enter a Model ID. ST falls back to the text field whenever the endpoint's model list can't be fetched; if the endpoint itself works, Bypass API status check silences the warning.
What does it cost?
Per-model, account-scoped — live pricing; top-up is a flat 1 RMB = $1 of credit on every payment method.
Try DaoXE — and benchmark it yourself
One key for GPT, Claude, Gemini, DeepSeek and more. Point the open benchmark at us and compare — don't take our word for it.