Document Translation

Document Translation — Sarvam AI

Live in-chat document translation → sarvam-mcp. This skill = SDK code.

Translating short text (a sentence or paragraph)? → text-translation — client.text.translate(), not this API.

[!IMPORTANT] Auth: api-subscription-key header — NOT Authorization: Bearer. Base URL: https://api.sarvam.ai (NOT /v1 — that prefix is only for the OpenAI-compatible chat endpoint)

What this API does

Translates a whole document (PDF, Word, Excel, PowerPoint, HTML) into up to 12 Indian languages per job, preserving layout and formatting. The output is a translated file in the original format — not plain text.

This is an asynchronous API. There is no single-call "translate this file" method. Every integration follows five steps: create → upload → start → poll → export.

Supported file formats

Quick Start (Python)

pythonCopy code
import os
from pathlib import Path

import httpx
from sarvamai import SarvamAI

client = SarvamAI(api_subscription_key=os.environ["SARVAM_API_KEY"])
document = Path("report.pdf")

# 1. Create job
job = client.document_translation.create(
    source_language_code="en-IN",
    target_language_codes=["hi-IN", "ta-IN"],
    original_filename=document.name,
)

# 2. Upload to the signed URL (SDK does NOT do this — use httpx/requests)
with document.open("rb") as f:
    httpx.put(
        job.upload_url,
        content=f.read(),
        headers={
            "Content-Type": "application/pdf",
            "x-ms-blob-type": "BlockBlob",
        },
        timeout=120.0,
    )

# 3. Start the pipeline
client.document_translation.start(job_id=job.job_id)

# 4. Poll live-status (every 10-15s) — see references/api-flow.md for full loop
# 5. Export each completed language — see references/api-flow.md
print(job.job_id)

Quick Start (JavaScript/TypeScript)

typescriptCopy code
import fs from "fs";
import { SarvamAIClient } from "sarvamai";

const client = new SarvamAIClient({
    apiSubscriptionKey: process.env.SARVAM_API_KEY!,
});
const documentPath = "report.pdf";

// 1. Create job
const job = await client.documentTranslation.create({
    source_language_code: "en-IN",
    target_language_codes: ["hi-IN", "ta-IN"],
    original_filename: documentPath,
});

// 2. Upload to the signed URL (SDK does NOT do this — use fetch)
await fetch(job.upload_url, {
    method: "PUT",
    headers: {
        "Content-Type": "application/pdf",
        "x-ms-blob-type": "BlockBlob",
    },
    body: fs.readFileSync(documentPath),
});

// 3. Start the pipeline
await client.documentTranslation.start(job.job_id);

// 4. Poll live-status (every 10-15s) — see references/api-flow.md for full loop
// 5. Export each completed language — see references/api-flow.md
console.log(job.job_id);

SDK method reference

Gotchas

Full Docs

Fetch supported formats, genre options, style guidelines, and language codes from:

• https://docs.sarvam.ai/llms.txt — comprehensive docs index
• Document Translation Overview
• Poll and Export Guide
• Rate Limits
• references/api-flow.md — full poll-and-export code loop