Nebula FAQs

We are rolling out Nebula approval in batches daily - stay tuned!

During Alpha, we are primarily testing the t0 model while conducting limited trials of the upcoming t1 model. The t0 model utilizes a mixture-of-agents architecture. We are targeting an early Q2 launch for t1, and will provide additional details around launch.

Nebula supports reading and writing capabilities on any EVM-compatible chain. View the chainlist for all supported chains.

Yes, you may ask Nebula questions in any spoken language.

• Define the Role Clearly: Specify the character’s role and behavior in your prompt.

  Example, “You are an assistant embedded in a blockchain application, designed to provide insights and execute transactions without introducing yourself.”

• Avoid phrases that suggest personality traits or self-awareness: Use Contextual Instructions and add specific instructions to avoid unnecessary introductions.

  Example, “Respond directly to queries without stating your identity or greeting users.”

• Set Output Expectations: Guide the format of responses.

  Example, “Provide concise answers or actions based on the user’s input without additional commentary.”

• Preload Context: Use Nebula’s context filters to scope responses within a specific domain (e.g., blockchain data or transactions) and avoid generic replies.

• Session Handling: Utilize Nebula’s session handling feature to maintain relevant context throughout interactions, ensuring responses align with the established character behavior.

We recommend using context filters to improve relevant responses. You may also modify hyperparameters such as temperature, top-p, and more using the OpenAI API. Learn more about our OpenAI API integration.

Nebula supports the read and write context of any verified contract or any contract on the thirdweb dashboard. For deploying through Nebula, supported contracts include Token - ERC20, NFT Collection - ERC721, Edition - ERC1155, and Split contracts.

If you have a published contract you would like to enable for deployment through Nebula, please contact us.

Nebula retains memory within the confines of a session. Learn more about sessions.

The context size or window is 128k tokens or approximately 96,000 words which may vary depending on the specific language and text characteristics.

Nebula is not currently open source. We are exploring open sourcing Nebula in the future.

Depending on the use case, Nebula can be used in a stateless manner by not reusing the same session. Nebula learns from each session, so having more contextual requests will be beneficial and are recommended.

Pricing options will be available in beta. Nebula is still being tested in alpha.