Use DaoXE in Crush

Crush runs agentic coding sessions in the terminal: it edits files, runs commands and reads your project's context files. Its config is a Bash file, and the provider path for non-OpenAI endpoints is the openai-compat type — exactly DaoXE's shape (cheapest APIs guide).

Why route Crush through DaoXE#

• Agent-grade coding, your models. Crush's planner, edits and tool calls all ride the provider you register — one DaoXE key covers hundreds of models from ~25 providers.
• Config is a Bash file. crushrc runs like .bashrc at startup, so keys can come from a password manager instead of sitting in JSON.
• Large and small model slots. A strong model for the main loop and a cheap one for background tasks map straight onto DaoXE's catalogue.

Set it up in Crush#

1. Register the provider. In ~/.config/crush/crushrc: provider add daoxe --type openai-compat --base-url "https://daoxe.com/v1" --api-key "${DAOXE_API_KEY:?set DAOXE_API_KEY}" — the docs' own openai-compat example uses a /v1-ending base URL and a shell-expanded key.
2. Register a model. model add daoxe/YOUR_EXACT_MODEL_ID with the exact ID from GET /v1/models, plus --context-window and optionally --can-reason / --supports-images — these capability flags decide what Crush asks the model to do.
3. Fill the model slots. model large daoxe/YOUR_EXACT_MODEL_ID for the main loop; model small daoxe/YOUR_SMALL_MODEL_ID for background work. References are always provider/model — the form printed by crush models.
4. Start a session. Run crush in your project; the TUI lists the provider's models. crush models prints the current large/small selections.

# ~/.config/crush/crushrc — register DaoXE as an openai-compat provider
provider add daoxe \
  --type openai-compat \
  --base-url "https://daoxe.com/v1" \
  --api-key "${DAOXE_API_KEY:?set DAOXE_API_KEY}"

# register a model, then put it in the large (coding) slot
model add daoxe/YOUR_EXACT_MODEL_ID --name "DaoXE primary" --context-window 128000
model large daoxe/YOUR_EXACT_MODEL_ID

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 agent, 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."}]
  }'

Frequently asked questions#