TokenDos API Reference: Dual Endpoints and Per-Model Reasoning Shapes

Endpoints and authentication

Both surfaces authenticate with a TokenDos API key: Authorization: Bearer <key> on the OpenAI surface, x-api-key: <key> on the Anthropic surface. Create keys in the dashboard; never commit them or paste them into logs.

Request examples

curl https://api.tokendos.com/v1/chat/completions \
  -H "Authorization: Bearer $TOKENDOS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-opus-5-5","messages":[{"role":"user","content":"ping"}],"max_tokens":64}'

curl https://api.tokendos.com/anthropic/v1/messages \
  -H "x-api-key: $TOKENDOS_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-opus-5-5","max_tokens":64,"messages":[{"role":"user","content":"ping"}]}'

Both are real billable calls; use a currently listed model ID from the marketplace. Common parameters (temperature, max_tokens, stream, tools, ...) follow each protocol's standard semantics; reasoning control fields vary per model as listed below.

Per-model reasoning shapes (OpenAI-compatible /v1 surface)

Column meanings: toggle = reasoning can be turned on/off; effort = enumerated ladder (none means it can be disabled); budget = thinking-token range. On the /anthropic surface the same models use native Anthropic thinking block semantics. Shapes are aligned per-model with upstream vendor definitions, snapshot 2026-10-22; models not listed expose no reasoning control on this surface.

Listing status and live prices follow the the marketplace and live pricing; the machine-readable snapshot is at transit-snapshot。This table is a static reference for integrations (coding clients, model catalogs, gateways) to verify parameter shapes, not a listing guarantee.