Transient Intelligence

Transient Intelligence — Quick start

Get your API key and complete your first successful analysis via the Transient Intelligence API.

1. Get Your API Key

Every request requires an API key in the x-api-key header. Create an account, then generate a key from your developer dashboard.

Step 1

Create an account

Create account

Step 2

Already have an account? Go straight to your dashboard.

Step 3

Generate your API key

Go to the Developers section and click "Create API key".

Get API key

Keep it safe. Your API key is tied to your account credits. Never commit it to version control or expose it client-side.

2. Your First Request

The simplest way to use the API is the one-call orchestrator endpoint. It handles upload, processing, and query answering in a single request.

Production endpoint

The examples below use the canonical production base URL: https://api.transientintelligence.com.

# Ask a question with inline data
curl -X POST "https://api.transientintelligence.com/api/models/v1/answer" \
-H "x-api-key: YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"question": "What was the Q3 revenue?",
"input": "The Q3 revenue reached $1.2M, driven by new enterprise signups.",
"top_k": 3
}'

// Ask a question with inline data
const response = await fetch("https://api.transientintelligence.com/api/models/v1/answer", {
method: "POST",
headers: {
"x-api-key": "YOUR_API_KEY",
"Content-Type": "application/json"
},
body: JSON.stringify({
question: "What was the Q3 revenue?",
input: "The Q3 revenue reached $1.2M, driven by new enterprise signups.",
top_k: 3
})
});
const data = await response.json();
console.log(data.citations);

# Ask a question with inline data
import requests

response = requests.post(
"https://api.transientintelligence.com/api/models/v1/answer",
headers={
"x-api-key": "YOUR_API_KEY",
"Content-Type": "application/json"
},
json={
"question": "What was the Q3 revenue?",
"input": "The Q3 revenue reached $1.2M, driven by new enterprise signups.",
"top_k": 3
}
)

data = response.json()
print(data["citations"])

# Evidence-backed answer
{
"success": true,
"sessionId": "session_abc123",
"citations": [
{
"doc_id": "inline_input",
"snippet": "The Q3 revenue reached $1.2M",
"locator": "paragraph 1"
}
],
"factual_confidence": {
"label": "Strong",
"message": "Answer is fully supported by evidence."
}
}

Let AI build your integration

Copy this prompt and paste it directly into Claude, ChatGPT, or any capable model. Fill in the placeholder at the bottom and it will generate working integration code for your stack.

Never paste your real API key into an AI chat

The prompt below tells the model to use an environment variable as a placeholder. Once you have the generated code, set TI_API_KEY in your environment — never hardcode it in the source.

You are helping me integrate the Transient (TI) API into my project.
## What TI does
TI is an evidence-first document analysis API. You send a question and document text;
it returns a cited answer grounded only in the document — no hallucination.
## API contract
Endpoint:  POST https://api.transientintelligence.com/api/models/v1/answer
Auth:      x-api-key request header — value read from TI_API_KEY environment variable
Body:      { question: string, input: string, top_k?: number }
Response:  { answer: { summary, confidence }, citations: [{ quote, locator, relevance }], run_id }
## Rules to follow in the code you generate
- Read the API key from the TI_API_KEY environment variable — never hardcode it
- Always check citations.length > 0 before using the answer
- If citations is empty, surface 'No evidence found' — do not fall back to answer.summary
- Pass document text as input, never a file path
- Set Content-Type: application/json on every request
## My integration target
[Replace this line: describe your stack, platform, and use case — e.g.
 'n8n workflow triggered by a new Google Drive file, post cited results to Slack'
 'Python Flask endpoint that accepts a PDF upload and returns a cited summary'
 'Make scenario: Notion page updated → analyse → write answer back to Notion']

The model will generate the full integration — HTTP calls, error handling, and empty-citations branching included. Come back to the AI Clients or Automation Platforms pages for platform-specific notes and system prompt templates.

Prompt your AI to use TI as a skill

Once the integration is live, your AI needs a system prompt that defines when to call TI and how to handle what it returns. Without it, the model will improvise — mixing retrieved evidence with its own training knowledge.

Prompt Library

The Prompt Library has the full operating prompt, client-specific system prompts for Claude Desktop, ChatGPT, Gemini and Grok, and use-case addons for legal, financial, and research documents.

Browse the Prompt Library →

SDKs & client libraries

There is currently no official SDK. The API is plain REST — any HTTP client works. The examples above are all you need to get started.

Auto-generate a client from the OpenAPI spec

Download the OpenAPI YAML and use openapi-generator or openapi-ts to generate a typed client in your language of choice.

Next Steps

Once you have the basic flow working, dive into the detailed integration contracts:

API Integration Basics →MCP Integration Quickstart →AI Client setup guides →Automation platform guides →