Quickstart

Create an API key, find a model, quote a request and follow its result.

1. Create an API key#

Create a key for your server. Permissions apply to the account that owns it.

Keep keys on your server. Start with read-only permissions and add only what your integration needs.

BASH

export TONG_API_KEY="YOUR_TONG_API_KEY"

Paid tools also require credits:spend. Delete permissions are separate.

2. Discover a model#

Replace example model IDs, reference fields and capability revisions with values from current discovery. Reference support varies by model and workflow.

curl -X POST 'https://platform.tongai.io/v1/tools/tong_workflows/operations' \  -H "Authorization: Bearer $TONG_API_KEY" \  -H "Content-Type: application/json" \  -d '{  "input": {    "query": {      "operation": "TEXT_TO_VIDEO"    }  }}'

Configuration

curl --version

Model catalogDiscover available models, capabilities and supported configurations.Model policiesReview each model’s published moderation policy before integrating.

3. Quote the configuration#

Quote first, then submit the exact quoted model and configuration. Poll the returned operation until it succeeds, fails, or needs your review.

curl -X POST 'https://platform.tongai.io/v1/tools/tong_generation_quote/operations' \  -H "Authorization: Bearer $TONG_API_KEY" \  -H "Content-Type: application/json" \  -H "Idempotency-Key: 45be69f1-c266-4a09-a82c-bf3c6303eb03" \  -d '{  "input": {    "body": {      "operation": "TEXT_TO_IMAGE",      "model": "tong_ea55aff4f3b5c8aba3a35569",      "routingParameters": {        "ratio": "16:9"      }    }  }}'

Configuration

curl --version

Use the quote submission value returned for this exact configuration.

Quote a generationGet a quote for the exact model, inputs and configuration before submitting the action.

4. Submit the request#

Examples use the public endpoint. Replace sample IDs with your own.

curl -X POST 'https://platform.tongai.io/v1/tools/tong_generation_submit/operations' \  -H "Authorization: Bearer $TONG_API_KEY" \  -H "Content-Type: application/json" \  -H "Idempotency-Key: 45be69f1-c266-4a09-a82c-bf3c6303eb03" \  -d '{  "input": {    "params": {      "id": "000000000000000000000001"    },    "body": {      "quoteId": "fixture-image-quote-token",      "prompt": "A quiet street at blue hour",      "operation": "TEXT_TO_IMAGE",      "inputAssetIds": [],      "parameters": {        "textToImageModel": "tong_ea55aff4f3b5c8aba3a35569",        "ratio": "16:9"      }    }  }}'

Configuration

curl --version

Keep one UUID for each intended action and reuse it after a timeout or retry. A different input with the same UUID is rejected.

5. Follow the result#

An accepted request is not a finished generation. Track the saved job until its output is ready.

curl 'https://platform.tongai.io/v1/operations/e8aa0a87-6cd9-42eb-975a-b72bf43218c8' \  -H "Authorization: Bearer $TONG_API_KEY"

Configuration

curl --version

JSON

{  "id": "e8aa0a87-6cd9-42eb-975a-b72bf43218c8",  "tool": "tong_generation_submit",  "category": "generation",  "channel": "api",  "state": "queued",  "status": "accepted",  "stage": "admission",  "dispatch_status": "pending",  "created_at": "2026-10-22T09:00:00.000Z",  "updated_at": "2026-10-22T09:00:00.000Z",  "detail_expires_at": null,  "detail_expired": false,  "timing": {    "dispatch_started_at": null,    "acknowledged_at": null,    "completed_at": null,    "dispatch_acknowledgement_ms": null,    "end_to_end_ms": null  },  "review": null,  "result": null,  "error": null,  "billing": {    "status": "pending",    "usage_credits": "0.000000",    "wallet_debited_credits": "0.000000"  }}

Requests and logsAn accepted request is not a finished generation. Track the saved job until its output is ready.Errors and rate limitsUnderstand error responses, retry safely and recover from failed generations.Credits and billingFollow quoted prices, actual usage and settled customer credits.