This track is for AI start-ups and companies building AI agents. Who should apply:

• Existing teams who are committed to building a long-term AI or AI x Crypto startup.
• Teams that already host or have started active development of a consumer-facing AI agent for which they intend to get mass distribution.

This track is for early-stage builders who want 1:1 support, including compute and API credits, for building high-quality agents. Feel free to tackle a Sentient request for an agent or come up with your own 💡 idea—we encourage both!

Ideal for unincorporated teams and students 🎓

• 💵 Compute Credits – Cloud allowance and API subsidies.
• 🎯 Mentorship – Access to Sentient engineers, BD, marketing, and product teams.
• 🏗️ Integration Guidance – Best practices for using the Sentient stack.
• 🏃‍♂️ First Row Seats - Fast Track to future Sentient Builder initiatives.

1. Sentient Agent API - preferred format that allows a rich set of intermediate events to be sent to the client (e.g. Sentient Chat) before the final answer. This format gives you the most precise control over the users’ experience and enables context-sharing between different agents on Sentient Chat.
2. OpenAI Chat Completions API - flexible format for any agent but does not expose intermediate agent events to the end-user. This format also enables context-sharing between different agents on Sentient Chat.
3. OpenAI Completions API (legacy) - simple format that requires you to store chat history internally to maintain context between follow-up questions. Using this format also makes your agent unable to share conversation context with other agents, thus your agent would not be used in multi-agent chats effectively.

We highly recommend that you implement streaming regardless of the API format you choose. Streaming would provide your end users with the best experience while waiting for their complete answer.