Enter the future of localization with Hera AI, InText’s automated post-editing, LQA and QE tool that takes context into account.
Hera AI improves machine translation quality and reduces localization costs by combining the speed and scalability of machine translation with the contextual intelligence of LLMs.
Automatically post-edits machine-translated bilingual files based on reference materials — including translation memories, glossaries, and style guides — to ensure consistent, high-quality output.
Reviews bilingual files against customizable LQA rules defined for a particular project or language pair, ensuring compliance with your quality standards. Operates in comment mode, leaving the original translations unchanged.
Evaluates bilingual files according to TAUS Post-Editing Guidelines and identifies which segments require light, full or no post‑editing. Operates in comment mode, leaving the original translations unchanged.
Generates translations directly from the source file into the target language while maintaining high quality through contextual understanding across multiple segments.
Language service providers (LSPs)
Speed up post-editing, LQA, and QE, reduce costs on linguists’ work, and handle even low-budget requests profitably without sacrificing quality.
Linguists
Improve the quality of MT output before editing, get instant QE and LQA feedback, so you can work faster, reduce repetitive fixes, and take on more projects.
Enterprises
Optimize your localization team’s productivity while cutting costs and without compromising on quality.
The overall impression is very positive. Context-aware AI-editing leads to a clearly noticeable improvement in quality.
Hera AI has cut our time per segment in half. Overall, it’s been a very positive experience. With this kind of text preparation and non-technical subject matter, it takes me on average 40 minutes for every 1000 words. I think it’s awesome.
Read more useful articles on our blog >>>
Right now, we are actively developing new features for Hera AI. Upcoming releases will include:
We welcome your suggestions for features or improvements that could make Hera AI more useful in your work.
Your data security is our priority. Hera AI does not store your input data or use it for training purposes.
When working with OpenAI, your uploaded files are transmitted directly from your device to OpenAI for processing. OpenAI guarantees that content submitted through its API is not stored, logged, or used to train models. All processing happens in a secure, one-time interaction, ensuring your content remains fully confidential.
When working with Ollama, your data never leaves your device. Ollama is an open-source tool that runs on your local machine and serves as a runtime for local LLMs — models like Llama, Mistral, or others you choose to deploy. Hera AI connects to Ollama directly on your device, meaning all processing happens entirely within your own infrastructure, with no data transmitted to any external server.
Token costs (charged by OpenAI per API usage) are not included in the plan prices.
Hera AI works directly with bilingual file formats, allows you to integrate reference materials, and applies context across multiple segments (not just one at a time). It also identifies logical inconsistencies in the source text and takes them into account, reducing the risk of errors.
With Hera AI, your data remains fully confidential. Hera AI is a desktop application. It does not store files on InText servers, and the app itself does not retain or reuse your content for training.
When you process segments using a cloud AI provider like OpenAI, they’re sent directly from your computer to the AI provider via a secure API connection for one‑time processing. The provider doesn’t store, log, or learn from your content — your translations stay private.
When you use Ollama with a local LLM, your data never leaves your device at all — processing happens entirely within your own infrastructure.
Yes. Hera AI supports local LLM integration via Ollama — an open‑source tool that runs on your machine and acts as a local server for models like Llama, Mistral, and others. Once Ollama is installed, running, and loaded with your chosen model, you connect it to Hera AI through the app settings. All processing happens locally via Ollama — no data is sent to any external server. This option is especially suited for projects with strict confidentiality requirements, such as legal, medical, or financial content.
Hera AI processes combined segments in a single request to the LLM. This is faster, more efficient, and more cost-effective than processing segments individually with large context windows.
Hera AI works with all languages supported by OpenAI. Output quality is highest and most predictable for widely used languages (particularly West European, East Asian, and Middle Eastern languages) with large training datasets. The system also performs well across Central European, Balkan, Baltic, and Caucasian languages.
For less common languages and rare combinations, we recommend using English as a pivot language between source and target to improve output consistency.
Hera AI currently supports .sdlxliff, .mqxliff, .mxliff, .xliff, .xlf bilingual file formats from Trados, memoQ, Phrase, Crowdin, Lokalise, Smartcat, Smartling, Wordbee, and XTM. Note that Hera AI works with exported bilingual files and does not connect directly to CAT tools.
Absolutely. Thanks to its broad contextual understanding, Hera AI recognizes incorrectly broken segments and adjusts the translation accordingly.
Yes. Hera AI applies your existing translation assets across the number of segments you define, improving consistency and reducing the need for manual editing. You can import them in .doc, .docx, or .txt formats.
Unlike basic machine translation tools, Hera AI processes bilingual files, applies project‑wide context across batches of segments, and supports reference materials, resulting in more consistent and accurate output. Hera AI doesn’t replace machine translation — it improves it. You can also choose your AI provider: connect to OpenAI cloud models or run local LLMs via Ollama for full control over your data.
We provide ready‑to‑use prompts for using Hera AI in different scenarios. You can use them as‑is or customize them to fit your specific workflows. To modify the prompt, simply use the Update button and describe what you’d like to achieve in plain language. Hera AI will then refine the default prompt for you.
In our experience, standard machine translation tools alone don’t meet the quality expectations of our clients, while human experts are typically reserved for high‑priority or sensitive content. To strike the right balance, we’ve developed Hera AI, which ensures both high quality and cost‑efficiency.
Shape the future of localization by filling out the first form below, downloading Hera AI, and starting a 30‑day free trial. After filling out the second form, you will also be able to get both free and paid plans of Hera AI.
Want to see Hera AI in action? Schedule a demo, or email us at hera.ai@intext.com.
For technical questions or support, you can reach us at hera.ai@intext.com or at the dedicated support forum.
