API Reference

Endpoint reference

Use these endpoints to build against the APIAny gateway. Core routes follow OpenAI-compatible conventions, while compatibility routes preserve old clients from the previous API platform.

Authentication

Protected endpoints accept Authorization: Bearer sk-your-api-key. During migration, x-api-key is also accepted, and Gemini-compatible routes accept x-goog-api-key.

Authenticated request

curl https://devnextapiany.mexxxxai.win/api/v1/models \
  -H "Authorization: Bearer sk-your-api-key"

Endpoints

GET	/api/v1/models	List active public models with type, capability, pricing, and parameters.
POST	/api/v1/chat/completions	Send OpenAI-compatible chat completion requests.
POST	/api/v1/images/upload	Upload JPEG, PNG, or WebP inputs with API key authentication.
POST	/api/v1/images/generate	Create an async image generation task.
POST	/api/v1/video/generate	Create an async video generation task.
GET	/api/v1/tasks/{task_id}	Poll an async media task by task id.
POST	/api/v1/api/generate	Accepts legacy media fields such as images, aspectRatio, imageSize, and video model aliases.
GET	/api/v1/api/result?id={task_id}	Legacy task result alias for clients using id as the query parameter.
GET	/api/v1/video/status?task_id={task_id}	Legacy video status response with success/data wrapper.
GET	/api/v1/internal/workflow-status?uuid={task_id}	Legacy workflow status alias for clients polling image task uuid values.
POST	/api/v1beta/models/{model}:generateContent	Accepts Gemini contents and generationConfig, then routes through the platform pipeline.
POST	/api/internal/jobs/process-pending-callbacks	Internal scheduler endpoint that retries pending async task callbacks.
POST	/api/internal/jobs/check-stale-media-tasks	Internal scheduler endpoint that marks stale processing media tasks failed and refunds reserved credits.
POST	/api/internal/jobs/check-api-channels	Internal scheduler endpoint that checks upstream channel health and updates routing status.
POST	/api/internal/jobs/scheduled-maintenance	Internal scheduler endpoint that rebuilds monthly summaries and deletes expired API detail rows.
GET	/api/internal/jobs/scheduled-maintenance	Returns recent scheduled-maintenance and retention cleanup runs.

Examples

Chat completion

curl https://devnextapiany.mexxxxai.win/api/v1/chat/completions \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "messages": [
      { "role": "user", "content": "Write a concise release note." }
    ],
    "max_tokens": 180
  }'

Media compatibility

curl https://devnextapiany.mexxxxai.win/api/v1/api/generate \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "videoai-3.1-fast",
    "type": "video",
    "prompt": "A product API network map, clean and cinematic",
    "aspectRatio": "16:9"
  }'

Gemini compatibility

curl https://devnextapiany.mexxxxai.win/api/v1beta/models/gpt-5.5:generateContent \
  -H "x-goog-api-key: sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [
      { "role": "user", "parts": [{ "text": "Explain API credit budgets." }] }
    ],
    "generationConfig": {
      "maxOutputTokens": 180
    }
  }'

Compatibility notes

API keys are hashed at rest and can be scoped by project, model allowlist, RPM, and credit budget.
Channels are selected by model channel groups, priority, and weight. Failed upstream attempts are recorded and retried on the next eligible channel for transient errors.
Run the channel health job from a scheduler to keep failover decisions fresh.
Run scheduled-maintenance daily to keep monthly revenue and cost rollups available and delete expired API detail rows after the retention window.
Media endpoints return task_id immediately; process pending media tasks from your scheduler and poll the task endpoint for completion.
Async media requests may include callback_url. It must be a public HTTPS URL; delivery failures are retried and do not change the task status.
Gemini-compatible streamGenerateContent requests are normalized through the non-streaming pipeline until streaming settlement is added.