Platform overview
Hello! Welcome to aiXplain, an agentic soltuions development and deployment platform. We offer a broad marketplace of standardized AI models, and provide tools to help you create custom AI solutions. You can interface with aiXplain in three ways:
- SDK,
- Bel Esprit,
- Studio.
What is aiXplain?
aiXplain is a platform that aims to simplify AI resources, bringing together models, datasets (soon) and metrics from both academic and commercial sources. They can be selected, combined, deployed and evaluated through different tools, and all managed using a Teams-based pay-as-you go Credits system.
Discover aiXplain's ready-to-use AI Models and Metrics.
Design and deploy custom Agents and Pipelines.
Benchmark models using Data and metrics.
FineTune pre-trained models using your data.
SDK
The aiXplain SDK is an open-source toolkit (Python, Swift). It enables you to access aiXplain's assets and tools in a standardized format. See below for the available features.
Get started with Python, Swift, or API Requests.
SDK Architecture
Studio
The aiXplain Studio provides a user interface to interact with aiXplain. It provides an easy way to
- Manage your aiXplain Teams and Credits
- Search and Try out our models in Discover
- Prototype pipelines in Design
- Access FineTune and Benchmark (including Bechmark reports)
See the Studio tour to learn more!
Bel Esprit
Bel Esprit is a Generative AI chat agent designed to simplify AI solution creation through natural language instructions. It selects AI models from aiXplain's marketplace, integrates them into deployable solutions, and offers user-friendly explanations. Bel Esprit connects users to AI specialists when further customization is needed.
Bel Esprit is currently integrated into Studio but will soon be available as a standalone application.
See our page on Bel Esprit to learn more!
Features Comparison
| Platform | SDK (Python) | SDK (Swift) | Notes | |
|---|---|---|---|---|
| Build and run agents | ❌ | ✅ | ❌ | |
| Run models, pipelines & metrics with single input | ✅ | ✅ | ✅ | |
| Run models, pipelines & metrics in batch mode | ❌ | ✅ | ✅ | |
| Run models, pipelines & metrics on datasets | ❌ | ✅ | ❌ | |
| Search assets | ✅ | ✅ | ❌ | Search parameters differ between Studio and SDK |
| Build pipelines | ✅ | ❌ | ❌ | Coming soon to the Python SDK |
| Onboard models (custom & Hugging Face) | ❌ | ✅ | ❌ | |
| Onboard datasets | ✅ | ✅ | ❌ | |
| Benchmark | ✅ | ✅ | ❌ | |
| FineTune | ✅ | ✅ | ❌ | |
| AutoMode | ❌ | ❌ | ❌ | Currently disabled |
| Build agents | ❌ | ❌ | ❌ | Coming soon! ✨ |
Feedback
Please share your feedback to help us improve! 🙏
The feedback button is at the bottom right of the Platform. Alternatively, message us on Discord.
Examples
- Bug report: [This] models don't work!
- Feature request: Please enable [this] use case.
- Improvement suggestion: Please make [this] easier to do.
- Technical support: I'm stuck trying to do [this]! Help!
