DaoXETypingMind is a bring-your-own-key frontend: chats, settings and API keys stay in your browser, and requests go straight to the provider you name. The Add Custom Model dialog is the generic OpenAI-compatible surface — exactly DaoXE's shape (cheapest APIs guide).
Why route TypingMind through DaoXE#
- Keys never leave the browser. TypingMind's docs state data and API keys "are stored locally in your browser and never sent to TypingMind's servers".
- Works on the web app and the static self-host build. The custom-model flow is the same in both, so DaoXE works even on a private deploy.
- Per-model control. Context length, plugins and vision toggles are set per custom model — you decide what each DaoXE model can do.
Set it up in TypingMind#
- Open the dialog. Click the Models icon in the left sidebar, then + Add Custom Model (top right) — the same flow TypingMind's docs use for Ollama, LM Studio and LocalAI.
- Set the API type. Choose the OpenAI API type (the docs' Gemini example shows the API Type selector; pick the OpenAI flavor for DaoXE).
- Paste the full endpoint. Endpoint:
https://daoxe.com/v1/chat/completions. TypingMind's documented examples all carry the full path (e.g.http://localhost:1234/v1/chat/completionsfor LM Studio) — not a bare base URL. - Model ID, headers, Test. Enter the exact model ID from
GET /v1/models, set the context length, add a custom header row (Authorization:Bearer sk-…) for the key, then click Test — the app checks the endpoint's capabilities before you save.
Caveats worth knowing
Endpoint is a full path: every documented custom-model example ends in /chat/completions; a base URL alone will fail the Test step. Model IDs are manual: there is no model-list fetch for custom models — copy exact IDs from GET /v1/models, one model per entry. Headers carry the key: the docs route custom keys through the Custom Headers row (their Gemini example uses apiKey); for OpenAI-shaped endpoints use Authorization: Bearer. Browser origin matters: requests go from your browser directly to the endpoint, so a self-hosted TypingMind on http://localhost can hit CORS walls that the hosted app (https) does not — if the Test passes in one and fails in the other, that's the difference. Test before save: the button checks the endpoint and detects capabilities; a wrong ID or path shows up there, not in chat. 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 the app, 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 TypingMind use DaoXE?
Yes — Models → Add Custom Model → OpenAI API type, endpoint https://daoxe.com/v1/chat/completions, key in a custom header.
Where do I put the API key?
In the Custom Headers row — Authorization : Bearer sk-your-daoxe-key. TypingMind documents custom headers for its Gemini custom model; the OpenAI type reads the same header convention.
Why does the model list not fill itself?
Custom models in TypingMind are manual — the docs' Ollama/LM Studio/LocalAI pages all have you type the model ID. Copy exact IDs from GET /v1/models.
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.