API Documentation

Upload text content to be chunked, indexed, and available for queries.

Request Body

Response 200

{
  "status": "success",
  "chunks_count": 12,
  "source": "report.txt"
}

Upload a file directly using multipart form data.

Supported formats

curl -X POST https://api.wauldo.com/v1/upload/file \
  -H "Authorization: Bearer $TOKEN" \
  -F "file=@contract.txt"

Response 200

{
  "status": "success",
  "chunks_count": 24,
  "source": "contract.txt",
  "file_size": 15234
}

Ask a question against your uploaded documents. Returns a verified answer with sources, confidence score, and full audit trail.

Request Body

Response 200

{
  "answer": "The contract specifies a 2% monthly late payment fee (Section 4.2).",
  "sources": [
    {
      "content": "Section 4.2: Late payments incur a 2% monthly fee...",
      "score": 0.92,
      "source": "contract.txt"
    }
  ],
  "audit": {
    "confidence": 0.92,
    "confidence_label": "high",
    "grounded": true,
    "retrieval_path": "BM25Reranked",
    "model": "qwen/qwen3.5-flash",
    "latency_ms": 1420,
    "sources_used": 2,
    "sources_evaluated": 5
  }
}

OpenAI-compatible chat endpoint. Works with any OpenAI SDK. Supports streaming.

Request Body

Returns available models. OpenAI SDK compatible.

curl https://api.wauldo.com/v1/models \
  -H "Authorization: Bearer $TOKEN"

List all document collections for the authenticated tenant.

Delete a collection and all its chunks. Useful for re-uploading updated documents.

Returns API health, RAG chunk count, Redis status, active provider, and uptime. No auth required.

{
  "status": "ok",
  "rag_chunks": 142,
  "redis": "connected",
  "provider": "openrouter",
  "uptime_seconds": 86400
}

Audit fields

# Show a trust badge based on confidence
audit = response["audit"]

if audit["grounded"] and audit["confidence_label"] == "high":
    show_badge("Verified", color="green")      # Safe to display
elif audit["confidence_label"] == "medium":
    show_badge("Likely correct", color="yellow") # Show with caveat
else:
    show_badge("Low confidence", color="red")    # Warn the user

# Log for monitoring
log(model=audit["model"], latency=audit["latency_ms"], path=audit["retrieval_path"])

pip install wauldo

from wauldo import WauldoClient

client = WauldoClient("https://api.wauldo.com")
client.login("demo", "demo_password")

# Upload a document
client.rag_upload("Your document text...")

# Query with debug info
result = client.rag_query("What are the key points?", debug=True)

print(result.answer)
print(result.audit.confidence)        # 0.92
print(result.audit.grounded)          # True
print(result.audit.retrieval_path)    # "BM25Reranked"
print(result.sources)                 # [Source(...)]

# Chat (OpenAI-compatible)
reply = client.chat_simple("Explain quantum computing")

# Conversation with memory
conv = client.conversation(system="You are a helpful assistant")
conv.say("Hello!")
conv.say("What did I just say?")  # remembers context

npm install wauldo

import { WauldoClient } from 'wauldo';

const client = new WauldoClient('https://api.wauldo.com');
await client.login('demo', 'demo_password');

// Upload & query
await client.ragUpload('Your document text...');
const result = await client.ragQuery('What are the key points?', 5, { debug: true });

console.log(result.answer);
console.log(result.audit.confidence);      // 0.92
console.log(result.audit.grounded);        // true
console.log(result.audit.retrievalPath);   // "BM25Reranked"

// Streaming chat
await client.chatStream(
  [{ role: 'user', content: 'Hello' }],
  { onToken: (t) => process.stdout.write(t) }
);

cargo add wauldo

use wauldo::WauldoClient;

let client = WauldoClient::new("https://api.wauldo.com");
client.login("demo", "demo_password").await?;

// Upload & query with debug
client.rag_upload("Your document text...").await?;
let result = client.rag_query_debug("What are the key points?").await?;

println!("{}", result.answer);
println!("Confidence: {}", result.confidence());  // 0.92
println!("Grounded: {}", result.grounded());      // true

// Streaming chat
client.chat_stream(messages, |token| {
    print!("{}", token);
}).await?;

Rate limits are per API key. Manage your subscription on RapidAPI.