> ## Documentation Index
> Fetch the complete documentation index at: https://api-docs.ollang.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Ollang Agentic Localization

> Ollang's multi-agent localization workflow — Agentic Localization and Agentic Turbo — for culturally aware translation across 100+ languages.

**Ollang Agentic Localization** is Ollang's own translation workflow and the default provider for every `translate` step. It is the only entry in the translation catalog that is a *pipeline* rather than a single model: agents are selected dynamically per segment based on content type, domain, and structural complexity, then their output is reconciled.

That design targets the gap between translation and localization. A frontier LLM will translate a marketing line accurately; a localization workflow decides whether the line should be translated at all, or replaced with something that lands the same way for a different audience.

<Info>
  **Available in Ollang Workflows** — Document Translation, Subtitle Translation, and AI Dubbing. See the [Translation catalog](/providers/translation) for the full comparison against 25 other engines.
</Info>

## Models

| Model                | Provider tag        | Order types                    | Status       |
| -------------------- | ------------------- | ------------------------------ | ------------ |
| Agentic Localization | `ollangAITranslate` | Document, Subtitle, AI Dubbing | **Default**  |
| Agentic Turbo        | `agentic-turbo`     | Document, Subtitle, AI Dubbing | Experimental |

## Key capabilities

* **End-to-end localization** across text, documents, subtitles, and dubbing scripts in **100+ languages**, from a single provider selection.
* **Dynamic multi-agent selection** — different agents handle different content, which is what carries quality through domain-specific and structurally complex material that a single prompt would flatten.
* **Culturally aware output** — idiom, register, and cultural reference are adapted rather than transliterated.
* **Format and structure preservation** across the document types Ollang ingests.
* **Steerable** through [Memory, glossaries, and custom instructions](/memory-guidelines-custom-instructions), which apply to the whole workflow rather than to one model call.

## Agentic Turbo

Agentic Turbo is the same shape of workflow tuned for throughput, while retaining contextual quality that pure machine translation does not reach.

It suits high-volume work — large subtitle backlogs, bulk document sets, and any pipeline where turnaround determines scope.

<Warning>
  Agentic Turbo is experimental. Benchmark it against your current provider in a separate Folder before moving a production Global Workflow onto it.
</Warning>

## Where it fits in a workflow

Agentic Localization is the default because most Orders are judged on localization quality rather than translation accuracy alone. It earns its place on content where the output is customer-facing and the phrasing carries value.

A common configuration is Agentic Localization as the Global Workflow default, with Agentic Turbo or [DeepL](/apis/translation-apis/deepl) overridden at Folder level for bulk content.

## Trade-offs

* **Latency** — a multi-agent pass does more work than a single model call, which shows up in turnaround on large batches.

## Reference

* [Translation provider catalog](/providers/translation)
* [Memory, guidelines, and custom instructions](/memory-guidelines-custom-instructions)
* [Workflows and provider architecture](/workflows-provider-architecture)