Open Source Startups

KeyDB is an open source, multithreaded database that serves as a faster alternative to Redis, designed for high-throughput applications requiring horizontal and vertical scaling. It supports multiple data structures, persistence options, and advanced features like MVCC non-blocking queries and cross-region multi-master replication.

2 investors

Khoj builds open-source AI software tools, including a research workbench, an AI co-worker platform, and a personal assistant agent application. The company targets researchers, knowledge workers, and individuals seeking transparent, locally-runnable AI tools.

2 investors

Lamin provides a lineage-native lakehouse platform for biological R&D that enables scientists to query, trace, and validate complex biological datasets across academia, biotech, and pharma. The platform integrates data versioning and cataloging capabilities to close the feedback loop between experimental results and machine learning models.

2 investors

LiteLLM provides an AI gateway that simplifies access to over 100 large language models through a unified OpenAI-compatible interface, with built-in features for fallbacks, spend tracking, and load balancing. The platform serves developers and engineering teams integrating multiple LLM providers into their applications.

2 investors

Magnitude is an AI coding agent optimized for open-source language models like GLM, Kimi, and DeepSeek, designed to generate and execute code with improved reliability and efficiency. The platform serves software developers and teams looking to build with open weights models at lower cost than frontier API pricing.

2 investors

Matano provides a cloud-native security information and event management platform that uses a Security Data Lake architecture as an alternative to traditional SIEM systems like Splunk or Elastic. The platform targets security teams operating in cloud-first environments.

2 investors

Mito is a spreadsheet-to-Python code generation tool that lets teams automate data workflows using a familiar spreadsheet interface within Jupyter environments. It generates production-ready Python code for data ingestion, transformation, and visualization without requiring users to learn new programming skills.

2 investors

Nao is an open-source analytics agent platform that enables users to query data warehouses in natural language through a managed or self-hosted chat interface. The platform is designed for data teams and business users to explore structured data, metrics, and models without writing SQL.

2 investors

OneCLI is an open-source credential gateway that secures AI agents by injecting API keys at the network layer, preventing agents and LLMs from accessing credentials directly. It works across any agent framework and enforces policies at the proxy level for endpoint blocking, rate limiting, and approval workflows.

2 investors

Understudy Labs offers an open-source toolkit that captures traces from LLM production workflows, benchmarks and A/B tests cheaper or open-weight models against them, and fine-tunes and deploys specialist models that users own and can host anywhere. It is aimed at developers and teams looking to migrate from proprietary LLMs to models they control.

2 investors

PeerDB provides fast and cost-effective data replication from PostgreSQL to data warehouses including Snowflake, BigQuery, and ClickHouse, as well as to event queues like Kafka and EventHubs. The open-source tool and commercial cloud service enable change data capture and parallel data migration for engineering and data teams.

2 investors

PipelineDB is an open-source PostgreSQL extension that continuously aggregates large streams of time-series data into summary data for realtime reporting use cases. It is aimed at analytics scenarios such as realtime dashboards that rely on summary data rather than exploratory analysis of raw datasets.

2 investors

RethinkDB is an open-source, scalable database for the realtime web that pushes updated query results to applications continuously instead of requiring polling. It provides a query language with features like table joins and groupBy, a parallelized architecture, and web-based cluster management tools for developers building realtime apps.

2 investors

Rivet provides infrastructure for AI agents and realtime applications, offering stateful compute primitives called Actors that run in Node.js or Bun backends. The platform supports deployment options ranging from local libraries to fully managed cloud services or self-hosted environments with compliance features like air-gapped and on-premises support.

2 investors

Stackup is a smart account platform that enables crypto companies to manage onchain financial operations through modern infrastructure for digital asset management. The platform targets enterprise-grade crypto businesses seeking to streamline transactions, hardware wallet management, and administrative processes.

2 investors

Tesseral is an open source platform that provides identity and access management for business software, with features including single sign-on, user provisioning, role-based access control, and audit logging. It serves organizations managing access for employees, customers, services, and AI agents.

2 investors

Tiptap is a headless, open-source editor framework that enables developers to build customized text editors with extensions for collaboration and AI capabilities. It serves developers and enterprise teams building applications that require rich text editing functionality.

2 investors

Tracer is a research lab building Echo, an inference system that coordinates open-weight models behind a single OpenAI-compatible endpoint for chat, code, and agent workloads. It focuses on allocating computation and coordinating complementary models to improve performance per token for teams running AI in production.

2 investors

Tusk is an AI agent that automatically generates unit tests for pull requests by understanding codebase and business context. The product helps software engineers save time on testing work.

2 investors

Metoro is an observability platform for Kubernetes environments that uses eBPF and AI agents to automatically detect production issues, investigate alerts, verify deployments, and generate fix pull requests. It requires no code changes or SDK integrations and operates at the kernel level to collect logs, metrics, traces, and profiling data.

2 investors

Zenbu is a hackable IDE built on Zenbu.js with native support for the Pi coding agent, enabling developers to customize the environment and run parallel development workflows with multi-agent support and git worktree integration.

2 investors

Activepieces is a no-code workflow automation platform that enables teams to connect apps, APIs, and webhooks to automate business processes across finance, marketing, sales, operations, and other functions. The platform offers AI-powered agents, enterprise integrations, role-based access control, and flexible deployment options for organizations of all sizes.

1 investor

Acyclic Labs runs agent swarms that break a task into parallel work across many sandboxed agents and merge their outputs into a single result, cutting the wall-clock time of long agentic workflows. Accessible via CLI, dashboard, or coding agents over MCP, it supports user-supplied model keys and self-hosted sandboxes atop an open-source harness, serving coding, research, and data workloads.

1 investor

Arroyo builds a stream processing engine designed for cloud environments and ease of use, enabling real-time data processing for companies building real-time applications. The platform serves organizations needing to process data with low latency without the need for specialized streaming infrastructure teams.

1 investor