> ## Documentation Index > Fetch the complete documentation index at: https://docs.sqwish.ai/llms.txt > Use this file to discover all available pages before exploring further. # Keep a family warm > Keep the model a family serves loaded between requests, or stop. Switching on a family that serves no model gets 409 family_empty, and one that is switched off 409 family_off. With `served`, the id of the model it served when the caller looked, only while it still serves that model: 409 served_changed otherwise. Keep warm follows the family when it moves. `switch_off` stops keep warm on families or models in the same step and leaves their switches on, so a request at the limit can swap. Past the account's limit, 409 warm_limit lists the models still kept warm after that, least recently used first (`used_at`), each with the keep warms that hold it: its own (`keep_warm`) and its families' (`families`). A size with no kept-warm place left answers 503 warm_capacity, unless `switch_off` stops a model of that size. ## OpenAPI ````yaml /openapi.json post /v1/families/{family}/keep-warm openapi: 3.1.0 info: title: DecisionOne workbench description: >- Native decisions, versioned datasets, adaptation and shared-base deployment. Authenticate with `Authorization: Bearer d1_sk_...` when accounts are on. A short guide for coding agents is at /llms.txt. Hosted account APIs accept a Bearer API key or a browser session. An API key reads its account but can't change it: keys, billing and settings change in the console. Browser mutations require the configured Origin and X-CSRF-Token from /auth/session. Account recovery and history remain available to blocked accounts. Waitlisted accounts can read their account summary and model catalogue and request early access; private member APIs require admission. Money fields ending in _usd are exact decimal strings; never interpret them as integer cents. Authenticated customer reads and existing-operation replays share read_rpm, and other customer writes except decisions and estimates share write_rpm. Exhausted budgets return 429 with Retry-After; unavailable admission storage returns 503. version: 0.1.0 servers: - url: https://console.sqwish.ai security: [] tags: - name: Decisions description: Probabilities for questions with fixed outcomes. - name: Models description: Base models, adapters and named models with versions. - name: Families description: >- Lines of models trained on one dataset, and the switches that let them answer calls. - name: Datasets description: >- Validated training examples. A dataset grows in versions, and a version never changes. - name: Evaluations description: Reusable model results on a dataset's fixed evaluation rows. - name: Fine-tuning description: Training jobs, their events and held-out results. - name: Prompt tuning description: Better wording for a decision, tested on your rows. Words, not weights. - name: Labeling description: Teacher labels for real queries, reviewed by you. - name: Accounts description: Email OTP sign-in, revocable sessions, security and API keys. - name: Billing description: Exact USD balances, purchases, automatic top-ups and referrals. - name: Integrations description: Endpoints other tools call, such as Claude Code. - name: System description: Health and discovery. paths: /v1/families/{family}/keep-warm: post: tags: - Families summary: Keep a family warm description: >- Keep the model a family serves loaded between requests, or stop. Switching on a family that serves no model gets 409 family_empty, and one that is switched off 409 family_off. With `served`, the id of the model it served when the caller looked, only while it still serves that model: 409 served_changed otherwise. Keep warm follows the family when it moves. `switch_off` stops keep warm on families or models in the same step and leaves their switches on, so a request at the limit can swap. Past the account's limit, 409 warm_limit lists the models still kept warm after that, least recently used first (`used_at`), each with the keep warms that hold it: its own (`keep_warm`) and its families' (`families`). A size with no kept-warm place left answers 503 warm_capacity, unless `switch_off` stops a model of that size. operationId: keepFamilyWarm parameters: - name: family in: path required: true schema: type: string title: Family requestBody: required: true content: application/json: schema: $ref: '#/components/schemas/KeepWarm' responses: '200': description: Successful Response content: application/json: schema: properties: continue: anyOf: - $ref: '#/components/schemas/Continuation' - type: 'null' description: >- What continuing it would do now, from the family's dataset and jobs. It doesn't check credit, storage allowance or queue room, which the start refuses with insufficient_credit, storage_allowance, queue_full or tenant_queue_full. Null for a family with no dataset. A list of families reads no dataset: there `new_data` and `retry` only say a version no finished job used is waiting, and the counts and fee are null. The answers of a switch, keep warm or rollback leave it out. id: type: string title: Id object: type: string const: family title: Object switched_on: type: boolean title: Switched On description: Whether calls to its name are answered. used_at: type: number title: Used At description: >- Its idle clock: the latest of its last production call by an API key that named it, its last evaluation that named it, and when it was switched on. switches_off_at: anyOf: - type: number - type: 'null' title: Switches Off At description: >- When idling switches it off unless it's used. Null while it is off or kept warm on its own switch, and when idling never switches anything off. A family's is also null while it serves no model. off_reason: anyOf: - type: string enum: - owner - idle - type: 'null' title: Off Reason description: >- Why it is off: owner when it was switched off by hand, idle when nothing used it for too long. Null otherwise. off_at: anyOf: - type: number - type: 'null' title: Off At description: When it went off, with off_reason. Null otherwise. dataset_id: anyOf: - type: string - type: 'null' title: Dataset Id description: The dataset it grows on. origin: anyOf: - type: string - type: 'null' title: Origin description: 'Where it started: a base model, or one of your models.' created_at: type: number title: Created At description: When it was named, as its first fine-tune was queued. served: anyOf: - $ref: '#/components/schemas/ModelVersion' - type: 'null' description: >- The model calls to its name get while it is switched on, or null for none yet. latest: anyOf: - $ref: '#/components/schemas/ModelVersion' - type: 'null' description: Its newest model, neither deleted nor discarded. models: items: anyOf: - $ref: '#/components/schemas/ModelVersion' - $ref: '#/components/schemas/DeletedModelVersion' type: array title: Models description: >- Every model of its line, oldest first. A deleted one shows only where it stood. keep_warm: type: boolean title: Keep Warm keep_warm_window: anyOf: - $ref: '#/components/schemas/KeepWarmWindow' - type: 'null' description: >- The hour of keep warm being billed for the model the family serves, while the family is kept warm. Null otherwise: a model kept warm on its own shows its own hour. keep_warm_stopped_at: anyOf: - type: number - type: 'null' title: Keep Warm Stopped At description: >- When keep warm switched off because the next hour couldn't be reserved, while it stays off. Otherwise null. move: anyOf: - $ref: '#/components/schemas/FamilyMove' - type: 'null' description: >- The move to its latest model it waits to make, and what it waits for. Null while it serves its latest model, or is switched off. moved: anyOf: - $ref: '#/components/schemas/FamilyMoved' - type: 'null' description: >- When it last moved by itself to the model it serves, and from which. Null when it never did, or serves another model or was rolled back since. jobs: items: $ref: '#/components/schemas/FamilyJob' type: array title: Jobs description: >- Its fine-tunes, newest first, with how each ended. Lists of families, and the answers of a switch, keep warm or rollback, leave it out. type: object required: - id - object - switched_on - used_at - switches_off_at - off_reason - off_at - dataset_id - origin - created_at - served - latest - models - keep_warm - keep_warm_window - keep_warm_stopped_at - move - moved title: Family 4XX: description: >- Client error. On a 422, `error.details` lists each field that failed. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' 5XX: description: Server error. A 503 names what is not running on this server. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' security: - BearerAuth: [] components: schemas: KeepWarm: properties: enabled: type: boolean title: Enabled max_price_usd: anyOf: - type: string pattern: ^\d{1,9}(\.\d{1,9})?$ - type: 'null' title: Max Price Usd description: >- The most you agree to pay for the first hour, which switching keep warm on holds, in USD, such as the price you were shown. If it costs more when it is admitted, it is refused with price_changed and nothing is held. served: anyOf: - type: string maxLength: 140 - type: 'null' title: Served description: >- The id of the model the family served when the caller looked: refused unless it still does. Null for none. switch_off: items: type: string type: array maxItems: 20 title: Switch Off description: >- Families, such as router, and models, such as router@2, whose keep warm stops in the same step, to make room when the account is at its limit (warm_limit) or the model's size has no kept-warm place left (warm_capacity). Their switches stay on. additionalProperties: false type: object required: - enabled title: KeepWarm Continuation: properties: state: type: string enum: - new_data - retry - no_new_data - not_enough_evaluation_data - queued - training - origin_gone - dataset_deleted - dataset_archived - fine_tuning_off title: State description: >- `new_data` when it can start on a version no finished job of the family used, and `retry` when it can start on the version its last run failed or was cancelled on. Otherwise what refuses it now: no training rows newer than the family last trained on, fewer evaluation decisions than its gate needs, a fine-tune of the family queued or training, an empty family whose origin model was deleted or discarded, its dataset deleted or archived, or fine-tuning off on this server. A family whose dataset was deleted never continues again: start a new family from its latest model. version: type: integer title: Version description: 'The dataset version a continuation trains on: its newest.' added_at: anyOf: - type: number - type: 'null' title: Added At description: When that version was added. added_from: anyOf: - type: string - type: 'null' title: Added From description: The family whose page added that version, or null when none did. evaluation_decisions: anyOf: - type: integer - type: 'null' title: Evaluation Decisions description: >- The evaluation decisions about new contexts its gate could judge. Null while a job is queued or training, and when it is refused whatever the data. needed: type: integer title: Needed description: How many of those the gate needs. rows: anyOf: - $ref: '#/components/schemas/ContinueRows' - type: 'null' description: The rows it would take. Null when evaluation_decisions is. price_usd: anyOf: - type: string - type: 'null' title: Price Usd description: >- The exact fee a continuation would hold now, as USD. Null unless it can start. type: object required: - state - version - added_at - added_from - evaluation_decisions - needed - rows - price_usd title: Continuation ModelVersion: properties: version: type: integer title: Version ref: type: string title: Ref adapter: type: string title: Adapter status: type: string title: Status discarded_at: type: number title: Discarded At on_at: anyOf: - type: number - type: 'null' title: On At description: >- When its own switch went on, so it answers calls to its number. Null while off. keep_warm: type: boolean title: Keep Warm description: Whether it is kept warm on its own switch. keep_warm_stopped_at: anyOf: - type: number - type: 'null' title: Keep Warm Stopped At description: >- When its own keep warm switched off because the next hour couldn't be reserved, while it stays off. Otherwise null. used_at: anyOf: - type: number - type: 'null' title: Used At description: >- Its idle clock: the latest of its last production call by an API key that named it, its last evaluation that named it, and when it was switched on. switches_off_at: anyOf: - type: number - type: 'null' title: Switches Off At description: >- When idling switches it off unless it's used. Null while it is off or kept warm on its own switch, and when idling never switches anything off. A family's is also null while it serves no model. off_reason: anyOf: - type: string enum: - owner - idle - type: 'null' title: Off Reason description: >- Why it is off: owner when it was switched off by hand, idle when nothing used it for too long. Null otherwise. off_at: anyOf: - type: number - type: 'null' title: Off At description: When it went off, with off_reason. Null otherwise. state: anyOf: - type: string enum: - loaded - loading - asleep - type: 'null' title: State parent: anyOf: - type: string - type: 'null' title: Parent created_at: anyOf: - type: number - type: 'null' title: Created At gate: anyOf: - $ref: '#/components/schemas/Gate' - type: 'null' heldout: additionalProperties: type: number type: object title: Heldout type: object required: - version - ref - adapter - status - on_at - keep_warm - keep_warm_stopped_at - used_at - switches_off_at - off_reason - off_at - state - parent - created_at - gate - heldout title: ModelVersion DeletedModelVersion: properties: version: type: integer title: Version ref: type: string title: Ref adapter: type: string title: Adapter parent: anyOf: - type: string - type: 'null' title: Parent description: The id of the model it trained from, or null from a base. status: type: string const: deleted title: Status deleted_at: type: number title: Deleted At type: object required: - version - ref - adapter - parent - status - deleted_at title: DeletedModelVersion description: >- A model you deleted, as its family's line shows it: where it stood, and when it went. KeepWarmWindow: properties: started_at: type: number title: Started At ends_at: type: number title: Ends At description: When this hour's hold ends and the next is reserved. reserved_usd: type: string title: Reserved Usd description: >- The hour's fee held, as exact USD. The minutes used are charged when the hour rolls over or keep warm is switched off. type: object required: - started_at - ends_at - reserved_usd title: KeepWarmWindow FamilyMove: properties: to: type: string title: To description: The model it moves to, its latest, such as router@5. waiting_for: type: string enum: - place - warm_place - warm_capacity - engines title: Waiting For description: >- `place` while the account has no place left for the move, which takes one only if the model the family serves keeps answering by its own switch, or it serves none yet. The move waits until a model is switched off or the tier goes up. A kept-warm family takes a warm place only if the model it serves stays warm by its own keep warm. `warm_place` while the account keeps as many models warm as its tier allows. That wait ends when a model's keep warm is turned off or the tier goes up. `warm_capacity` while the model's size has no kept-warm place free on our engines. That wait ends when one frees. Turning the family's keep warm off ends either wait. `engines` otherwise: the model loads, or waits to be tried again after a failed load. reason: anyOf: - type: string enum: - live_limit - warm_limit - warm_capacity - type: 'null' title: Reason description: >- The code of the refusal that holds a waiting move: `live_limit` for `place`, `warm_limit` for `warm_place` and `warm_capacity` for `warm_capacity`, as switching on, keep warm and rollback name them. Null while it waits for engines. type: object required: - to - waiting_for - reason title: FamilyMove FamilyMoved: properties: at: type: number title: At description: When it moved. from: anyOf: - type: string - type: 'null' title: From description: >- The model it served before, such as router@4, or null when it served none. type: object required: - at - from title: FamilyMoved FamilyJob: properties: id: type: string title: Id status: type: string enum: - queued - running - cancelling - succeeded - failed - cancelled title: Status outcome: anyOf: - type: string enum: - joined - gate_failed - failed - cancelled - rolled_back - type: 'null' title: Outcome description: >- How it ended: its model joined the family, failed its gate, the run failed or was cancelled, or a rollback discarded the model it made. Null while it runs. reason: anyOf: - type: string - type: 'null' title: Reason description: >- Why it made no model, in one line: the gate's first reason, or the first line of the error of a run that failed. Otherwise null. dataset_version: type: integer title: Dataset Version output_model: anyOf: - type: string - type: 'null' title: Output Model created_at: type: number title: Created At price_usd: type: string title: Price Usd description: The fee held for this job, as exact USD. charge: type: string enum: - free - reserved - charged - released title: Charge description: >- What the fee is doing: held while the job runs, charged once training finishes, whatever the gate says, released if the run fails or is cancelled. type: object required: - id - status - outcome - reason - dataset_version - output_model - created_at - price_usd - charge title: FamilyJob ErrorResponse: description: >- The body of every error. app.failure() builds it; this model documents it. properties: error: $ref: '#/components/schemas/ErrorBody' required: - error title: ErrorResponse type: object ContinueRows: properties: new: type: integer title: New description: Training rows added since the family last trained. earlier: type: integer title: Earlier description: Earlier training rows sampled to train on with them. evaluation: type: integer title: Evaluation description: Evaluation rows its gate would measure on. type: object required: - new - earlier - evaluation title: ContinueRows Gate: properties: passed: type: boolean title: Passed against: type: string enum: - parent - base title: Against reasons: items: type: string type: array title: Reasons groups: additionalProperties: additionalProperties: true type: object type: object title: Groups deltas: additionalProperties: type: number type: object title: Deltas ece: additionalProperties: anyOf: - type: number - type: 'null' type: object title: Ece type: object required: - passed - against - reasons title: Gate ErrorBody: properties: code: description: A stable code to branch on, such as invalid_request. title: Code type: string message: description: What went wrong and what to change, in plain words. title: Message type: string retryable: description: >- True when another attempt may succeed; it does not prove that no work was done. Respect Retry-After when present. Only operations that support Idempotency-Key replay a prior result, including /v1/decide for 24 hours. title: Retryable type: boolean request_id: description: Also sent as X-Request-ID. Quote it in bug reports. title: Request Id type: string details: anyOf: - items: $ref: '#/components/schemas/ErrorDetail' type: array - type: 'null' default: null description: On validation errors, one entry for each field that failed. title: Details required: - code - message - retryable - request_id title: ErrorBody type: object ErrorDetail: properties: path: description: Where the problem is, such as ["body", "decisions", 0, "outcomes"]. items: anyOf: - type: string - type: integer title: Path type: array message: title: Message type: string type: anyOf: - type: string - type: 'null' default: null title: Type required: - path - message title: ErrorDetail type: object securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: d1_sk_... description: Create an API key in the console. Required when accounts are enabled. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.