Search Companies
Find startups and other companies in the COSS ecosystem
Cloud platform for web development and deployment
Vercel is a platform for frontend developers, providing the frameworks, workflows, and infrastructure to build a faster, more personalized web. Their core product is built on technologies like JavaScript, React, and Next.js, and they serve businesses looking to improve their web performance and developer experience.
IDE for parallel AI coding agents
Orca provides a desktop integrated development environment for running multiple AI coding agents in parallel. It combines isolated Git worktrees, agent terminals, browser tabs, diff review, and Git integration for local or remote developer workflows. The open-source Orca project is available under the MIT License for desktop, mobile, and remote environments.
Context compression for AI agents
Headroom Labs develops open-source context-compression software for AI agents. Its tools reduce repeated content in tool calls, database rows, file reads, and RAG payloads before model inference. The product is available as an agent wrapper, proxy, SDK, and MCP server, with enterprise VPC deployment and support options.
AI company developing large language models.
DeepSeek is an AI company focused on developing and applying advanced artificial intelligence technologies. Their core product is built on large language models and aims to serve enterprise clients. They specialize in tools and solutions for various industries.
COSS Weekly Newsletter
Stay up to date with the latest news, funding rounds, and announcements from the COSS universe.
Check out COSS Weekly on the webEnterprise AI gateway for LLMs and agents
TrueFoundry provides an enterprise AI gateway for governing access to LLMs, MCP servers, and agents. The platform combines gateway controls with LLM operations, model serving, tracing, and observability capabilities for enterprise teams. Its MIT-licensed TrueForge project is an open-source agent harness that connects models, tools, MCP servers, sandboxes, and infrastructure.
Tabular data prediction models
Synthefy develops foundation models for structured data. Its open-source Nori project provides training-free regression and predictive distributions for tabular data. The company also offers a managed API for deploying these models.
AI and agent engineering platform
Arize AI provides an AI and agent engineering platform for building, observing, and evaluating AI applications. It helps close the loop between AI development and production.
OpenTelemetry-native observability platform
Dash0 is an OpenTelemetry-native observability platform that provides unified monitoring of logs, metrics, and traces for developers and SREs. Built on open standards including OpenTelemetry, PromQL, and Perses, it offers infrastructure monitoring, distributed tracing, log management, Kubernetes monitoring, and AI-powered SRE agents with transparent, usage-based pricing.
Continuous profiling for performance & costs
Polar Signals Cloud offers always-on, zero-instrumentation continuous profiling for CPU, GPU, and Memory. This helps organizations improve software performance, understand incidents, and significantly lower infrastructure costs by identifying and optimizing inefficient code paths.
Platform for durable execution of code.
Temporal provides an open-source platform for durable execution, enabling developers to build applications that are resilient to failures. It ensures that workflows can be paused, resumed, and retried without losing state.
Unified AI compute platform with MAX & Mojo
Modular is an AI infrastructure company that builds a unified compute platform for developing and deploying generative AI applications. Its products include MAX, a high-performance AI inference engine, and Mojo, a Python-compatible programming language designed for high-performance AI and systems programming on CPUs and GPUs.
Unified API for large language models (LLMs)
OpenRouter provides a platform for accessing and using large language models (LLMs) through a unified API. It focuses on simplifying the integration of various LLMs for enterprises, enabling them to leverage AI technologies effectively.
Postgres sync engine for local-first apps
ElectricSQL provides a data platform for building collaborative, multi-agent systems using Postgres sync. The platform offers a read-path sync engine that handles partial replication, data delivery, and fan-out from Postgres to client applications. It includes PGlite, an embeddable Postgres with real-time reactive bindings, and Durable Streams, a data primitive for the agent loop. ElectricSQL offers both a self-hosted open-source version and a managed cloud platform.
Data intelligence platform for AI solutions.
Databricks is a cloud-based data, analytics, and AI platform designed to help enterprises build, scale, and govern their data infrastructure. It pioneered the data lakehouse architecture, which combines the capabilities of data warehouses and data lakes for unified analytics. It serves enterprise customers.
AI agent testing and evaluation platform
LangWatch develops an AI-agent engineering platform for testing, evaluating, simulating, and monitoring LLM-powered applications. Its Apache-2.0 core can be self-hosted and supports developers building reliable agents.

