Reference

API reference

Every endpoint, generated from the OpenAPI 3.1 document. Send real requests from the console.

OpenAPI 3.1.0Version 1.0.0Live document

Venice-compatible routes are described at the envelope level: their bodies follow the Venice and OpenAI schemas. Tokentide-native routes are fully typed below.

Base URLs

https://api.tokentide.app/v1

The API is also served under the Venice-style alias /api/v1. See Base URL.

Authentication

API keyAuthorization: Bearer …: tt_live_… (production) or tt_test_… (preview/staging).

Request headers

Optional on every inference route.

Name	Type	Description
`X-Tokentide-Lane`header · optional	string	Lane selection. Low Tide is opt-in, non-streaming and best-effort.One of `standard`, `lowtide`
`X-Tokentide-Lowtide-Fallback`header · optional	string	Reroute to Standard when Low Tide is closed or capped.One of `standard`, `none`
`X-Tokentide-Max-Multiplier`header · optional	string	Never charge above this multiplier of the Venice list price. The API key's own maximum multiplier still applies: the lower of the two wins.Pattern `^0\.(?=\d*[1-9])\d{1,6}$\|^1(\.0{1,6})?$`
`X-Tokentide-Max-Cost-USD`header · optional	string	Reject at admission if the worst-case charge exceeds this amount.
`X-Tokentide-Fallback-Models`header · optional	string	Comma list (≤ 3) of opt-in fallback models.
`X-Tokentide-Debug`header · optional	string	Adds diagnostic response headers; nothing extra is stored.One of `1`
`Idempotency-Key`header · optional	string	Duplicate detection for 24 h. Results are never replayed.
`X-Tokentide-Upload`header · optional	string	Large Payload Lane reference: `<upload_id>; key=<base64>`.

Response headers

Priced responses carry what you paid and why. Every header, explained.

X-Request-Id
X-Tokentide-MultiplierLocked multiplier (decimal).
X-Tokentide-Lane
X-Tokentide-Epoch
X-Tokentide-Cross-Epoch
X-Tokentide-List-Cost-USDNon-streaming responses only.
X-Tokentide-Charge-USDNon-streaming responses only.
X-Tokentide-Model-Privacy
X-RateLimit-Limit-Requests
X-RateLimit-Remaining-Requests
X-RateLimit-Reset-Requests

Errors

Errors use the OpenAI-compatible envelope with a stable code. All error codes, with what to do.

Inference

POST/v1/chat/completions

Chat completions (streaming, E2EE passthrough)

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, text/event-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

Read the guide

POST/v1/responses

Responses API (typed SSE events)

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, text/event-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

Read the guide

POST/v1/embeddings

Embeddings

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

Read the guide

POST/v1/decisions

Typed judgments (Jev System One)

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/systemone

Alias of /decisions

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

Images

POST/v1/image/generate

Generate images (Venice-native)

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/images/generations

Generate images (OpenAI-compatible)

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/image/edit

Edit an image

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/image/multi-edit

Compose several images

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/image/upscale

Upscale an image

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/image/background-remove

Remove an image background

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

GET/v1/image/styles

Image style presets

Public

Request body

No request body.

Responses

200OK

Audio

POST/v1/audio/speech

Text to speech

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/audio/transcriptions

Speech to text

API key

Accepts the Tokentide request headers.

Request body

multipart/form-data

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

Not available in the console.Read the guide

POST/v1/audio/quote

Price a music or sound job

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/audio/queue

Queue a music or sound job (charged when queued)

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/audio/retrieve

Poll a music or sound job; returns the audio when ready

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/audio/complete

Delete a music or sound job's media

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

Video

POST/v1/video/quote

Price a video job

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/video/queue

Queue a video job (charged when queued)

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/video/retrieve

Poll a video job; returns the MP4 when ready

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json, application/octet-stream
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

POST/v1/video/complete

Delete a video job's media

API key

Accepts the Tokentide request headers.

Request body

application/json

Same body as the Venice endpoint (OpenAI-compatible where applicable), forwarded byte for byte.

Responses

200Upstream response passed through byte-for-byte (except the final usage/cost chunk on streams).application/json
402Refused before forwarding; nothing is charged. insufficient_balance: your spendable balance cannot cover the worst-case cost of the request. spend_limit_reached: a daily (UTC day), weekly (ISO week, from Monday 00:00 UTC) or monthly (UTC calendar month) spend cap on the API key or the account would be exceeded. When the cap is used up, the message names it and when it resets, and Retry-After gives the seconds until the reset; a cap of 0 blocks all spend and has no Retry-After. When only this request's worst-case cost does not fit in the room left, the message says so and Retry-After is 60 seconds (none when the request could cost more than the whole cap): lower max_completion_tokens or wait for requests in flight to finish. Error envelope (see Errors)application/json
ErrorError envelope (see Errors)application/json

Models

GET/v1/models

Models with API list prices and live Tokentide prices

Public

Request body

No request body.

Responses

200OK

Read the guide

GET/v1/models/traits

Model traits

Public

Request body

No request body.

Responses

200OK

GET/v1/models/compatibility_mapping

Model compatibility mapping

Public

Request body

No request body.

Responses

200OK

Market

GET/v1/market

Live multipliers, lane status, epoch state and tide tables

Public

Request body

No request body.

Responses

200OKapplication/json

Read the guide

GET/v1/market/stream

Server-sent price ticks every 15 s

Public

Request body

No request body.

Responses

200SSEtext/event-stream

Read the guide

GET/v1/market/rules

The live rules of the market: formulas, parameters, key times, Low Tide fences

Public

Parameters

Name	Type	Description
`format`query · optional	string	`json` (default) for the structured rules, `markdown` for the same rules as a written guide.One of `json`, `markdown`Default `json`

Request body

No request body.

Responses

200OKapplication/json, text/markdown
ErrorError envelope (see Errors)application/json

GET/v1/market/estimate

What a request costs now, through the day and at tomorrow's opening

Public

Parameters

Name	Type	Description
`model`query · required	string	Model id (or alias), as listed by GET /v1/models.From 1 to 128
`input_tokens`query · optional	integer	Prompt tokens, for token-priced models (text, embeddings, decisions).From 0 to 100000000
`output_tokens`query · optional	integer	Completion tokens, for token-priced models. Reasoning tokens count as output.From 0 to 100000000
`units`query · optional	string	Quantity in the pricing unit GET /v1/models lists for the model (`tokentide.standard.unit`): images, edits, upscales, seconds of audio, or millions of characters (3,500 characters = 0.0035).Pattern `^\d{1,9}(\.\d{1,6})?$`
`at`query · optional	string (date-time)	Also price the request at this time (ISO 8601) from the published tide table: today, or tomorrow once tomorrow's table is published at 12:00 UTC.

Request body

No request body.

Responses

200OKapplication/json
ErrorError envelope (see Errors)application/json

POST/v1/market/quote

Lock a multiplier for 60 s (never across the daily reset)

API key

Request body

application/json

QuoteRequest
Field	Type	Description
`lane`optional	string	One of `standard`, `lowtide`Default `standard`

Responses

200OK

Read the guide

Account

GET/v1/account

Balance, tier, limits, spend caps and Low Tide allowance

API key

Request body

No request body.

Responses

200Balance (paid, promotional, leased and spendable credit), spend since 00:00 UTC, trust tier, rate limits and the spend caps of the calling key and of the account (weekly_cap_usd and account_*_cap_usd included). Spend caps apply per UTC day, per ISO week (from Monday 00:00 UTC) and per UTC calendar month; null means no cap, and a cap of "0" blocks all spend until it is changed.application/json
ErrorError envelope (see Errors)application/json

Read the guide

GET/v1/account/usage

Request metadata (never content), as JSON or CSV

API key

Parameters

Name	Type	Description
`limit`query · optional	integer	Default `100`From 1 to 1000
`before`query · optional	string	Cursor: return rows older than this request id.Pattern `^req_[0-9A-HJKMNP-TV-Z]{26}$`
`from`query · optional	string (date-time)
`to`query · optional	string (date-time)
`model`query · optional	string
`endpoint`query · optional	string
`request_id`query · optional	string	A request id from the X-Request-Id response header.Pattern `^req_[0-9A-HJKMNP-TV-Z]{26}$`
`format`query · optional	string	One of `json`, `csv`Default `json`

Request body

No request body.

Responses

200One page of request metadata, newest first. JSON: data, has_more and next_before (pass it as before for the next page). An API key sees its own requests; format=csv returns the same columns as one CSV page, without a cursor.application/json, text/csv
ErrorError envelope (see Errors)application/json

Read the guide

Large Payload Lane

POST/v1/uploads

One-time private upload slot for client-side encrypted files

API key

Request body

application/json

UploadRequest
Field	Type	Description
`purpose`required	string	The endpoint the payload is for.One of `audio_transcription`, `image_edit`, `image_multi_edit`, `image_upscale`, `image_background_remove`
`bytes`required	integer	Ciphertext size in bytes (nonce + ciphertext + tag).From 0 to ∞
`sha256`required	string	Lower-case hex SHA-256 of the ciphertext.Pattern `^[0-9a-f]{64}$`
`filename`optional	string	From 1 to 200Pattern `^[\w.\- ]+$`
`content_type`optional	string	From 3 to 100Pattern `^[\w.+-]+\/[\w.+-]+$`

Responses

200OK

Read the guide

Service

GET/v1/status

Health and epoch state of the API

Public

Request body

No request body.

Responses

200OK

Read the guide

GET/v1/openapi.json

This OpenAPI document

Public

Request body

No request body.

Responses

200OK