Flue
Flue is an open agent framework for building durable AI agents with a programmable TypeScript harness. Created by the team behind Astro, it uses a React-like hooks API to compose agent behavior, persistent state, lifecycle events, models, tools, sandboxes, subagents, skills, and MCP servers in code. Agents are stateful and addressable over HTTP, keep context across conversations and events, and can change capabilities as work progresses. Flue records every session in a durable stream, so accepted work survives crashes, restarts, and deploys; interrupted sessions resume automatically and clients reconnect without starting over. Developers can run agents locally, in CI, from their own backend, or orchestrate them with systems such as Cloudflare Workflows and Inngest. Secure sandboxes let agents run commands, edit files, and do real work, while tools connect them to APIs and data. Flue is powered by Pi and supports different LLM providers, allowing teams to choose the models they want.
Learn more
Calljmp
Calljmp is a developer-first AI agent runtime designed to build, run, and scale long-running stateful workflows written in TypeScript. While many modern tools like Mastra AI provide rich frameworks to define agents and workflows, Calljmp focuses on actually running them reliably in production.
Calljmp combines agent logic, durable execution, human-in-the-loop pause/resume, retries with idempotency, and built-in observability into a unified execution environment. Developers implement agents as code, and the runtime guarantees reliable execution, state persistence, and operational visibility without gluing together custom queues, databases, and monitoring stacks.
Calljmp is ideal for engineering teams, product developers, and backend architects who want to embed intelligent agents into product systems while offloading execution complexity to a purpose-built runtime.
Learn more
epho
Epho turns coding agents into an API, letting developers run Claude Code, Codex, or OpenCode in isolated cloud sandboxes through a single HTTP endpoint. Send a prompt, choose a harness and model, attach repositories and files, connect MCP servers, and pass environment variables or provider credentials; Epho boots the environment, clones the code, wires in the tools, and streams the agent’s work back as it happens. Runs can return live events, tool calls, edits, final answers, and artifacts, or execute asynchronously with polling and webhooks. Chats are durable, so follow-up turns can resume with the same filesystem, checkout, agent session, system prompt, model, and MCP configuration even if the original sandbox is gone. Private GitHub, GitLab, and Bitbucket repositories are supported, and agents can read code, make changes, run tests, and iterate on failures just as they would locally. Every event is persisted, allowing interrupted streams to reconnect without losing the run.
Learn more
Hermes Agent
Hermes Agent by Nous Research is an open-source autonomous AI agent designed to run locally on your server and improve over time. It operates independently from traditional chatbots by maintaining persistent memory and learning from past interactions. The agent can integrate with multiple platforms such as Slack, Discord, Telegram, and WhatsApp through a unified gateway. Hermes supports automation tasks like scheduling reports, managing workflows, and executing commands using natural language. It also enables parallel task execution through subagents, improving efficiency for complex operations. With built-in tools for web browsing, search, and code execution, it provides a versatile environment for various tasks. Overall, Hermes Agent acts as a continuously evolving AI system that adapts to user needs and workflows.
Learn more
