How Baseten launched Baseten Hosted Tools and Baseten Grounded Inference on X
nference is everything
What Baseten launched

Until now, adding web search to open-source models meant hand-wiring orchestration, managing separate keys, and paying latency taxes on every round trip. Today, we're solving that. We're excited to introduce Baseten Hosted Tools and Baseten Grounded Inference to bring real-time https://t.co/NyOhrCtatt
How LaunchVideo.com rated the launch
LaunchVideo.com rating
About the launch
Baseten Hosted Tools and Baseten Grounded Inference add server-side capabilities to the open-source and custom models running on Baseten's inference platform, aimed at developers who have been stitching together their own tool loops to make models useful in production. The first capability shipping is web search, which gives models hosted on Baseten access to information that is as recent as the current day, specific, and absent from their training data . Instead of managing separate API keys and orchestration code, developers add a tool selector to a standard Chat Completions, Messages, or Responses request and Baseten runs the model and the search provider together in a server-side loop. At launch, Baseten is offering access to four web search providers: Exa, Keenable, Parallel, and You.com , with pricing passed through to the Baseten bill. The model decides whether a prompt needs web search based on its native tool use capability and may execute multiple iterations until it is confident , and teams can tune that behavior through a configurable iteration budget that scales from one-shot lookups to deeper research workflows. Baseten interleaves model inference and search requests in a server-side loop while streaming tool-use and tool-result blocks as live progress to the client , which cuts round-trip latency compared with client-side tool orchestration. The launch matters because open-source models like Llama, DeepSeek, and GLM are increasingly viable substitutes for closed frontier APIs, but they have lacked the batteries-included tool infrastructure that makes those APIs easy to ship against. Baseten was founded in 2019 in San Francisco by Amir Haghighat (CTO), Tuhin Srivastava (CEO), Pankaj Gupta, and Philip Howes (Chief Scientist), and this release extends the company's inference stack from raw model serving toward the agent and grounding layer that application teams have been building by hand.
Tags
Questions
Frequently asked questions
The Baseten launch post reached 31K views on X, with 101 likes, 12 reposts, 10 comments, 41 bookmarks.
Baseten launched on X on September 16, 2026 with the message "nference is everything".
Baseten Hosted Tools and Baseten Grounded Inference add server-side capabilities to the open-source and custom models running on Baseten's inference platform, aimed at developers who have been stitch.
LaunchVideo.com rated the Baseten launch 2.20 out of 5: hook 2/5, editing / creativity 2/5, copy 2/5, sentiment of launch 3/5, distribution strategy 2/5.







