Amazon Bedrock
Fine-tune Amazon Nova Micro, Lite or Pro in your AWS account.
Read the guide →PRODUCT · DEDICATED AI MODELS
Train your AI model in your own cloud
From a spreadsheet of your team’s past decisions to a model that’s yours, trained and answering in your own Amazon Bedrock, Microsoft Foundry or Google Cloud. StayCharted does every step: the training file, the training job, testing on rows it never saw, and putting it to work.
Start free How it works →What stays in your account
- The training file goes to your account for training and is deleted when training ends.
- The trained model stays in your account and answers there.
- Your cloud bills you for training and answers, at its own rates. Your StayCharted plan is separate.
- Bedrock uses a role created by a template. Foundry uses Microsoft sign-in and two roles on one resource. Google Cloud uses a service account and key you create.
StayCharted holds and processes your uploaded examples to prepare the file and test the model. See our Privacy Policy and security page.
Choose your cloud
Microsoft Foundry
Fine-tune GPT-4.1 mini, nano and more in your Foundry resource.
Read the guide →Google Cloud
Fine-tune Gemini in your Google Cloud project, even straight from a Google Sheet.
Read the guide →How they compare
| Feature | Amazon Bedrock | Microsoft Foundry | Google Cloud |
|---|---|---|---|
| Models you can train | Amazon Nova Micro, Lite, Pro | GPT-4.1 mini, GPT-4.1 nano, GPT-4.1, GPT-4o mini | Gemini 3.5 Flash; 2.5 models for eligible existing projects |
| Where | us-east-1 | Your Foundry resource’s region; Global training may process data elsewhere | Your tuning region, such as us-central1 |
| Connecting | A template creates a role; no key | Sign in with Microsoft; no key | A service account and its key |
| Testing before use | Held-out rows | Developer deployment: no hourly hosting fee; inference billable | Held-out rows |
| Answering cost | Per token, plus custom-model storage charges | Per token plus hourly hosting while published | Per token, at the applicable tuned-model rate |
| Training file | Your bucket; deleted after training | Your resource; deleted after training | Your bucket; deleted after training |
Availability and costs depend on the model, region and deployment. Each guide links to provider documentation and pricing.
Use your own AI, without training
Already use OpenAI, Claude, Gemini or a model on Bedrock or Foundry? Your AI Models ask it with your categories, your labeling guide and the most similar past examples in every question. Nothing is trained.
Use your own AI →Who it’s for
Training in your own cloud requires Business. It is a Dedicated AI Model, like one trained on StayCharted’s servers; only where it trains differs. Your AI Models are an add-on on Essentials and Business.
Compare plans →Questions about training in your own cloud
Can I fine-tune a model in my own AWS account?
Yes. StayCharted fine-tunes Amazon Nova on Amazon Bedrock in your AWS account, from a spreadsheet, and the custom model stays in your account.
Can I fine-tune in my own Azure or Microsoft Foundry?
Yes. StayCharted fine-tunes GPT-4.1 models in your Foundry resource. It tests on a Developer deployment with no hourly hosting fee; inference remains billable; the paid deployment is made when you publish.
Can I fine-tune Gemini in my own Google Cloud?
Yes, Gemini on Google Cloud (Vertex AI, now Agent Platform), including straight from a Google Sheet with the StayCharted add-on.
Does my data leave my cloud account?
The training file goes to your account and is deleted when training ends; the trained model stays and answers there. StayCharted holds your examples to build the file and test the model, as it does for any model.
Can I use my own OpenAI or Claude key instead?
Yes, with Your AI Models: StayCharted asks your AI with your categories and past examples. OpenAI and Claude aren't trained by StayCharted.
What does it cost?
Your cloud bills training and answers at its rates (Foundry adds an hourly hosting fee for a published model). StayCharted bills your plan, as for any model.
Amazon Bedrock, Amazon Nova and AWS are trademarks of Amazon.com, Inc. or its affiliates. Microsoft, Azure and Microsoft Foundry are trademarks of the Microsoft group of companies. Google Cloud, Gemini and Vertex AI are trademarks of Google LLC.
STAYCHARTED AI MODEL TRAINER
Train AI to categorize the way your team does.
Start with examples you already have, see how often it’s right on items it never saw, and use confidence to prioritize the ones a person should review.
Start free