OAK Just Got a Major Upgrade! GPT-4.1 Turbo, MCP & PDF Support, Smarter Knowledge Base & More!
Hey Product Hunt Community! 👋
Excited to share some big news from the OAK team! We've just pushed a significant update packed with features designed to make building and deploying AI agents even better, based heavily on user feedback and our vision for a more powerful, intuitive platform.
Here’s a breakdown of what’s new:
✨ Cleaner, Refreshed UI: We've polished the interface, focusing on the default tools and chat experience. It's now sleeker, more intuitive, and hopefully, a lot more pleasant to work with day-to-day.
🧠 Latest Model Support (incl. GPT-4.1 Turbo): By upgrading to the newest Vercel AI SDK, OAK now supports the latest and greatest models, including GPT-4.1 Turbo. This means access to more powerful reasoning, longer context windows, and the cutting edge of AI capabilities for your agents.
📄 PDF Support for OpenAI Models: This has been a popular request! You can now directly integrate and process PDF documents using any of the OpenAI models configured in OAK. Unlock the knowledge stored in your PDFs and make it accessible to your agents seamlessly.
🏷️ Smarter Knowledge Base with Tagging & Metadata: This is a big one for knowledge retrieval! You can now add tags and metadata to documents in your knowledge base. Why is this cool? It allows your agent to filter and retrieve information with much greater precision. Instead of just searching content, it can look for documents tagged 'Q1_Reports' or with metadata indicating a specific 'Project_ID'. This leads to far more relevant and context-aware responses.
⚙️ MCP Support Added: We're also rolling out support for MCP (Multi-Cue Prompting or relevant protocol), expanding OAK's integration capabilities and allowing for more sophisticated agent interactions and workflow integrations. We'll share more specific use cases and guides on this soon!
These updates represent a significant step forward in usability, intelligence, and knowledge management for OAK. We're striving to build the best platform for creating capable AI agents, and these features are core to that mission.
We'd love to hear your thoughts!
What do you think of the new features?
How do you see the tagging/metadata impacting your knowledge base use?
Any questions about the new model support or PDF integration?
Drop your feedback and questions below – we're eager to chat!
Cheers,
Replies