Skip to content

Atraium 0.2.1 Preview 1

Pre-release
Pre-release

Choose a tag to compare

@daemosofchaos daemosofchaos released this 31 Jul 18:10
· 5 commits to main since this release

Atraium 0.2.1 Preview 1

Atraium 0.2.1 Preview 1 expands your model choices. In addition to OpenAI-compatible services, Atraium now supports Anthropic Claude and Google Gemini across supported features.

Recommended target

  • Windows desktop: recommended preview experience
  • Android: not included in this release unless listed in the attached assets
  • Web: not currently supported

Highlights

More model providers

Choose the provider that best suits each part of your workflow.

  • Continue using OpenAI-compatible models and endpoints.
  • Use Anthropic Claude for Chat and Utility work.
  • Use Google Gemini for Chat, Utility, memory, images, voice, and live interactions.
  • Keep existing OpenAI-compatible settings: no migration is required.

Mix providers by feature

You are no longer limited to one model provider for the entire app.

For example, you can use:

Chat and Utility       → Anthropic Claude
Memory and search      → Google Gemini
Image generation       → Google Gemini
Realtime and voice     → Google Gemini

You can also keep OpenAI-compatible services for any supported feature. Each feature can have its own provider, model, endpoint, and API key.

Guided configuration improvements

The guided setup now makes provider choices clear from the start.

  • Choose a provider for Chat and Utility.
  • Select providers separately for Embeddings, Images, Realtime, Speech-to-text, and Text-to-speech.
  • See only the choices that are available for the feature you are configuring.
  • Use provider-specific API keys without replacing keys already configured for other services.
  • Start with suggested models, including Claude 4.8 and Gemini 3.5 Flash.

Capability overview

Feature Available providers
Chat and Utility OpenAI compatible, Anthropic Claude, Google Gemini
Memory and semantic search OpenAI compatible, Google Gemini
Image generation OpenAI compatible, Google Gemini
Realtime and voice OpenAI compatible, Google Gemini

Anthropic Claude is currently available for Chat and Utility. Atraium does not present Claude as an option for features that Anthropic does not provide.

What to try

  1. Open Settings or Guided configuration.
  2. Configure an Anthropic or Google AI API key.
  3. Choose a provider and model for Chat and Utility.
  4. Optionally choose different providers for memory, images, realtime, and voice.
  5. Start a new conversation and try your selected Chat model.

Important limitations

  • This is preview software and may contain breaking issues.
  • Provider availability, model access, regions, quotas, and feature support can vary by account.
  • OpenAI-compatible providers can differ in route and feature support.
  • Tool behaviour depends heavily on the selected model.
  • Cost controls are incomplete; monitor every configured provider account.
  • Powerful local workflows should be tested with non-critical files first.
  • Support is best-effort.

Downloads

This preview release includes:

  • Atraium-0.2.1-preview.1-win-x64-setup.exe
  • Atraium-0.2.1-preview.1-win-x64-portable.zip
  • Atraium-0.2.1-preview.1-SHA256SUMS.txt

Verification

Verify the SHA-256 checksum of the downloaded installer or archive against the attached checksum file.

Token usage, costs, and credentials

You are solely responsible for all usage and charges from the providers, endpoints, plugins, MCP servers, and infrastructure you configure.

Atraium does not include provider credentials, model credits, or token allowance. It does not enforce your provider budget, control provider pricing, or issue refunds for provider charges.

Using more than one provider can create charges across multiple accounts. Chat, memory, images, voice, realtime sessions, and tool-rich workflows may all be billed separately by the configured provider.

Read Token usage and provider billing before using paid credentials.

Feedback

Use GitHub Issues for reproducible bugs and structured feedback. Use Discussions for broader questions and workflow ideas.

Do not publish API keys, tokens, private endpoint URLs, confidential prompts, source code, or sensitive files.