​
Introduction

Postman is the industry-standard API testing platform. By importing the Nebula Unified API collection, you can explore and test all Nebula endpoints — chat completions, image generation, video generation, embeddings, reranking, and more — directly from a friendly GUI, with zero coding required.

​
Download Files

Download both files below and import them into Postman:

Nebula Unified API Collection

The full API collection with pre-built requests for every endpoint, organized into folders.

Nebula Environment

Pre-configured environment variables: base_url, api_key, and default model IDs for each capability.

​
Quick Setup

​
Step 1 — Import the Collection

  1. Open Postman and click Import (top-left).
  2. Drag and drop Nebula_Unified_API.postman_collection.json onto the import dialog, or use Upload Files.
  3. Click Import to confirm.

​
Step 2 — Import the Environment

  1. Click Import again.
  2. Upload Nebula_Unified_API.postman_environment.json.
  3. Click Import to confirm.

​
Step 3 — Set Your API Key

Your API key is not included in the environment file. You must add it manually to keep it private.
  1. In Postman, open the Environments panel (left sidebar).
  2. Select Nebula Unified API - Environment.
  3. Find the api_key variable row and enter your Nebula API key in the Current value column.
  4. Click Save (Ctrl/Cmd + S).
Always fill in Current value, not Initial value. Current value is local to your machine and is never synced to Postman’s cloud.

​
Step 4 — Activate the Environment

In the top-right corner of Postman, open the environment dropdown and select Nebula Unified API - Environment.

​
Step 5 — Send Your First Request

  1. Expand the Nebula Unified API collection in the left panel.
  2. Open folder 00 - Getting Started → click Smoke Test - Chat Completion.
  3. Click Send.
  4. You should receive a 200 OK with the message Nebula API is ready.

​
Collection Structure

The collection is organized into logical folders:
FolderWhat’s inside
00 - Getting StartedList Models, Smoke Test Chat
01 - Text / OpenAI-compatibleChat Completion, Streaming, Responses API, Web Search, Structured Output
02 - Native Provider APIsClaude Messages (incl. streaming & token count), Gemini generateContent (incl. streaming & image output)
03 - Embeddings & RetrievalOpenAI-compatible Embeddings, Rerank
04 - Image GenerationGemini, Doubao Seedream (text-to-image), Doubao Seededit (image edit), GPT Image, Qwen Image
05 - Video GenerationSubmit & poll tasks for Sora 2, Veo, Wanxiang, Doubao Seedance, and more
06 - Realtime & VoiceWebSocket / realtime voice endpoints
Provider model foldersAuto-generated folders grouped by provider and capability from /v1/models

​
Environment Variables

VariableDefaultDescription
base_urlhttps://llm.ai-nebula.comNebula API base URL — do not change
api_key(empty)Your Nebula API key — fill this in
default_chat_modelqwen-plusModel used in chat completion requests
default_responses_modelgpt-5.2Model used in Responses API requests
default_claude_modelclaude-sonnet-4-5-20250929Model used for Claude native endpoints
default_gemini_modelgemini-2.5-flashModel used for Gemini native endpoints
default_image_modelgemini-2.5-flash-imageDefault image generation model
default_video_modelveo-3.1-fast-generate-001Default video generation model
default_embedding_modeltext-embedding-3-smallDefault embedding model
default_rerank_modelrerank-v1Default rerank model
video_task_id(example)Auto-filled by video submit test scripts
image_base64(placeholder)Replace with a real base64 data URL for image-edit requests
image_urlhttps://example.com/image.jpgReplace with a real image URL for image-edit requests

​
Tips & Best Practices

You can override the environment model for any single request without changing the environment variable. In the request Body, simply replace {{default_chat_model}} with the exact model ID you want, e.g. claude-sonnet-4-6.
Postman supports Server-Sent Events (SSE). Open a streaming request (e.g. Chat Completion - Streaming) and look for the Visualize tab or the raw response body to see streamed chunks in real time.
Video generation is asynchronous:
  1. Send Submit Video Task — the test script auto-saves task_id to video_task_id.
  2. Send Query Video Task (uses {{video_task_id}}) — repeat until status is succeeded.
  3. Use the returned URL or call Download Video to retrieve your file.
For image-edit endpoints (e.g. Doubao Seededit, Qwen Image Edit):
  • Set image_base64 in the environment to a data:image/png;base64,... string, or
  • Set image_url to a publicly accessible image URL.
After a successful run, click Save as Example to capture the response. This makes it easy to share reference responses with your team.

​
Troubleshooting

  • Confirm the environment is selected (top-right dropdown).
  • Verify api_key has a value in Current value (not just Initial value).
  • Make sure your Nebula account has sufficient credit.
  • Check your internet connection.
  • Ensure base_url is exactly https://llm.ai-nebula.com with no trailing slash.
  • If you are behind a corporate firewall, configure a Postman proxy.
  • Some models require specific account permissions or are region-restricted.
  • Try switching to a different default_chat_model like qwen-plus, or use gpt-5.2 through the Responses API.
  • Run List Models in the 00 - Getting Started folder to see models available to your key.