Train the models only you can build. SF Tensor provides one optimized, cross-vendor training stack for enterprise post-training and frontier pre-training.
Product · How it works
- Founded
- 2025
- Team
- 6 people
Browser storage is unavailable. Keep this page open and export your collection before leaving.
Train the models only you can build. SF Tensor provides one optimized, cross-vendor training stack for enterprise post-training and frontier pre-training.
Product · How it works
Primitive is email infrastructure for AI agents: universal email addresses, hosted functions, and transactional email over the protocol every business already uses. It also ships a ChatGPT app — an in-agent MCP App UI built on the…
Butter is building bVisor, an incredibly lightweight sandbox runtime for AI agents.
Founders · Product
We research the psychology of how LLMs learn, and turn it into new training techniques that train custom models and harnesses together.
Founders · Product
We build distributed compute clusters with the cheapest CPUs and GPUs across Hyperscalers and Neoclouds for AI. Our mission is to bring frontier-grade infrastructure to everyone. We're starting by building large scale high…
Founders · Product
Inkbox gives AI agents an identity, communication channels, and their own computer.
Founders · Pricing · Product · How it works
TrainLoop makes it effortless for developers to supercharge LLM performance through reinforcement learning.
Founders
Turn videos into structured data with Dragoneye zero-shot vision models. No data labeling or model training.
Founders · Product · How it works
RunAnywhere makes the best open models faster and cheaper to run, for everyone. Wally serves frontier open models through one API key and one CLI, and our open-source SDKs run models on the device.
Founders · Product
Respan is an open-source AI gateway with built-in observability and evals.
Founders · Product · How it works
Wafer serves open models on a stack tuned to your real traffic, layer by layer: model, decoding, engine, kernels and hardware.
Product
Use Pakistan's best AI calling platform to automate appointment booking, lead qualification, and routine customer service. 100+ realistic Urdu voices. APIs for Text to Speech (tts). Try for free.
Founders · Origins · Product · How it works
We handle the infra so you can focus on building agents that search, scrape, and interact with the web at scale.
Founders · Product
Schedule training jobs, track experiments, and serve endpoints. Use one tool or all three, with reliability built into each.
Founders · Pricing · Product
Refresh builds simulation engines for coding and computer use, for frontier labs to train the next generation of virtual software engineering collaborators.
Founders · How it works
Notte: The fast, reliable browser framework for AI Agents. Simplifies LLM web tasks, handles CAPTCHAs & offers dev tools. Build smarter agents.
Customers · Product
AI lab building interaction models to make AI symbiotic with humans. Our early models sit at the Pareto frontier of intelligence and interactivity, measured by time to first audio token, and enable voice agents that are both highly…
Product
Cloudglue APIs turn videos into video context for AI — including speech, diarization, visual descriptions, sound, and more — so developers can search, chat, and build video-aware AI. The video understanding infrastructure for modern AI.
Founders · Product
GPU orchestration, model inference, and an agent platform. Adopt Paladin, Ion, or Talos independently, or connect the full Cumulus stack.
Founders · Product
Make your own Discord bot with camelBot. camelAI also makes open-source AI agent hosting and an open-source AI coding assistant.
Pricing
Texel is a platform that enables developers to generate hardware-accelerated APIs for running AI pipelines 10 times faster, including image and video applications. For example, developers can use the latest NVIDIA GPUs or Google TPUs…
Founders · Product · How it works
Compresr is an LLM context-compression API: send long context plus your query, get back a shorter context that keeps the answer-bearing tokens and drops the rest. Cut token cost and latency; at light compression, match or beat…
Founders
The coding agent router. Keep Claude Code, Codex, OpenCode or Pi, and Dari picks the model for each turn, aware of provider caches, behind one endpoint.
Product · How it works