InfluxDB 3 OSS is now GA. Transform, enrich, and act on time series data directly in the database. Automate critical tasks and eliminate the need to move data externally. Download now. Learn more →
Top 23 TypeScript Agent Projects
-
CopilotKit
React UI + elegant infrastructure for AI Copilots, AI chatbots, and in-app AI agents. The Agentic Frontend 🪁
Project mention: A2UI in Practice: Build Agent-to-User Interfaces Using A2A + AG-UI | dev.to | 2025-12-16In this guide, you will learn how to build full-stack agent-to-user interface (A2UI) agents using the agent-to-agent (A2A) protocol, the AG-UI protocol, and CopilotKit.
-
SurveyJS
SurveyJS: Build JSON-Driven Surveys and Forms with Full Data Control. Add the SurveyJS UI components to your JS app (React/Angular/Vue3). Securely collect and analyze data without sending it to 3rd-party servers. Fully customizable, works with any backend, and ideal for data-heavy apps. Learn more.
-
composio
Composio equips your AI agents & LLMs with 100+ high-quality integrations via function calling
Project mention: Composio equip AI agents with 100 quality integrations via function calling | news.ycombinator.com | 2025-06-27 -
we built an execution engine ourselves https://github.com/simstudioai/sim/tree/main/apps/sim/execut... and for the infra for background jobs, we use https://trigger.dev/
-
activepieces
AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents
- Project mention: Show HN: I built E2E Test Agent – describe tests in plain English,AI executes it | news.ycombinator.com | 2025-11-15
hey! cool project, feels very similar to [stagehand](https://github.com/browserbase/stagehand), although stagehand doesn't have much in the way of e2e testing. might be worth building on top of that though since playwright MCP can overwhelm an agent with tool overload
-
mastra
The TypeScript AI agent framework. ⚡ Assistants, RAG, observability. Supports any LLM: GPT-4, Claude, Gemini, Llama.
Project mention: Building PhishNet: An AI Cybersecurity Agent for Detecting Phishing Threats with Mastra | dev.to | 2025-11-04In this article, I’ll walk you through how I built PhishNet, an AI-powered phishing detection agent using Mastra — the AI framework that changes everything.
-
-
InfluxDB
InfluxDB – Built for High-Performance Time Series Workloads. InfluxDB 3 OSS is now GA. Transform, enrich, and act on time series data directly in the database. Automate critical tasks and eliminate the need to move data externally. Download now.
-
12-factor-agents
What are the principles we can use to build LLM-powered software that is actually good enough to put in the hands of production customers?
Project mention: Show HN: Radkit – Build reliable AI agents with Rust | news.ycombinator.com | 2025-11-13We (me and my co-founder) have been working on this library for a while and would love to get your thoughts.
github.com/agents-sh/radkit
radkit.rs/getting-started/
Most of the agent librararies out there just adds a huge prompt, put all the tool definitions and call the LLM on a loop hope and for the best.
This is not a way to build reliable ai agents. A better way to build ai agents is described in 12-factor agents by Dex Horthy here.
https://www.youtube.com/watch?v=8kMaTybvDUw
https://github.com/humanlayer/12-factor-agents
Another limitations of current agentic frameworks is their support for interoperability among agents is very limited.
Google came up with https://a2a-protocol.org/latest/ (now under linux foundation) to address this, but the agent frameworks trying to do as much as possible and keep the end users locked into their "ecosystem" instead of supporting such protocols.
Everyone is so focused on MCP for some reason. (no hate for mcp, radkit supports mcp as well).
We belive in an agentic future where businesses expose ai agents and not mcp servers or rest apis. If you build an agent using radkit, it is a2a-protocol compliant from the get go.
With all of this in mind, we are building `radkit`.
It is still early stages. But we have reached a milestone and wanted to share with the community.
-
-
screenpipe
AI app store powered by 24/7 desktop history. open source | 100% local | dev friendly | 24/7 screen, mic recording
https://github.com/mediar-ai/screenpipe is promising, however it has some issues with my setup. I'm personally just dumping all the data with ffmpeg + x11grab, will figure out what I want to do with it later
- Project mention: 👽 Extract Thousands of Rows of Data Without Writing Code (Open Source) | dev.to | 2025-07-17
Explore the project on GitHub: https://github.com/getmaxun/maxun
-
Looks like a great change! I'll take it for a spin in a moment.
I really like the "subagent" feature in Claude Code — it's super useful to manage context in complex codebases. Here are some examples of agents that can be useful: https://github.com/humanlayer/humanlayer/tree/main/.claude/a...
Would it make sense to have a similar feature in Codex CLI? I often do "spec-driven development", which is basically a loop of:
-
git-mcp
Put an end to code hallucinations! GitMCP is a free, open-source, remote MCP server for any GitHub project
Project mention: Here's how git-mcp.io shows the typewriter effect on its landing page. | dev.to | 2025-07-19This component renders the typewriter-effect in gitmcp.io
- Project mention: Multi-head latent attention and other KV cache tricks explained | news.ycombinator.com | 2025-01-28
Thanks! Yes, will push it as template to our repo (https://github.com/PySpur-Dev/pyspur) soon! We used o1 and claude3.5
-
genkit
Open-source framework for building AI-powered apps in JavaScript, Go, and Python, built and used in production by Google
Project mention: Show HN: I built an open source drag and drop editor for Genkit AI flows | news.ycombinator.com | 2025-09-13Hi,
I have been building small AI Agents for quite some time now using various frameworks and one thing that always bugged me was that iterating on small things like prompts, flows, tools etc always took a code change + deployment of the app. While the prompt part can be solved with Langfuse I haven't found a good way to keep the flow management remote (and open source).
Lately I have been working with Genkit (https://github.com/firebase/genkit) and love how modular it is. So I thought why not build a UI builder on top of that that can handle simple flows, prompts and basic tracing.
And here we are with a first early version: https://flowshapr.ai
-
- Project mention: Show HN: N8n-Style Actions and AI Agents in TypeScript | news.ycombinator.com | 2025-12-16
- Project mention: Show HN: Agent Heroes – Generate, animate and schedule your AI characters | news.ycombinator.com | 2025-04-08
-
-
-
lmnr
Laminar - open-source all-in-one platform for engineering AI products. Create data flywheel for your AI app. Traces, Evals, Datasets, Labels. YC S24.
Project mention: Show HN: We instrumented Claude Agent SDK using a tiny Rust proxy | news.ycombinator.com | 2025-12-04At Laminar (https://github.com/lmnr-ai/lmnr) we're building open source AI observability platform in Rust.
We obsess over instrumentation DX for our Python and TS SDKs and in this new blog we outline how we made the most seamless way of instrumenting recently released claude agent sdk
-
Most AI frameworks provide the ability to manage both chat history and memory as it's important to ensure the history is managed to ensure it fits within the context window alongside the other elements of context. Taking LangGraph as an example, short term memory is managed as part of agent state using a checkpointer, meanwhile long term memory is persisted to long term stores:
-
The original issue in claude code was closed with "wontfix", so here is a second take: https://github.com/anthropics/claude-code/issues/7060
Dancing with AGENTS/CLAUDE/windsurf files symlinks/imports becomes really annoying. Using projects like https://github.com/intellectronica/ruler for some reason doesn't stick (yet another tool to install, yet another cognitive debt to pay).
Please help bring attention to that issue.
-
Stream
Stream - Scalable APIs for Chat, Feeds, Moderation, & Video. Stream helps developers build engaging apps that scale to millions with performant and flexible Chat, Feeds, Moderation, and Video APIs and SDKs powered by a global edge network and enterprise-grade infrastructure.
NOTE: The open source projects on this list are ordered by number of github stars. The number of mentions indicates repo mentiontions in the last 12 Months or since we started tracking (Dec 2020).
TypeScript Agents discussion
TypeScript Agents related posts
-
Show HN: Thufir – Claude Code plugin to solve production issues
-
Agent Factory Recap: Can you do my shopping?
-
Show HN: N8n-Style Actions and AI Agents in TypeScript
-
Show HN: WillItFront.Page – Test your HN post before submitting
-
Show HN: SIM – Apache-2.0 n8n alternative
-
Show HN: SIM – Apache-2.0 n8n alternative
-
Show HN: We instrumented Claude Agent SDK using a tiny Rust proxy
- A note from our sponsor - InfluxDB www.influxdata.com | 22 Dec 2025
Index
What are some of the best open-source Agent projects in TypeScript? This list will help you:
| # | Project | Stars |
|---|---|---|
| 1 | CopilotKit | 27,029 |
| 2 | composio | 26,217 |
| 3 | sim | 23,694 |
| 4 | activepieces | 19,875 |
| 5 | stagehand | 19,513 |
| 6 | mastra | 18,875 |
| 7 | agentic | 18,067 |
| 8 | 12-factor-agents | 17,080 |
| 9 | ai-pdf-chatbot-langchain | 16,243 |
| 10 | screenpipe | 16,173 |
| 11 | maxun | 14,054 |
| 12 | humanlayer | 8,006 |
| 13 | git-mcp | 7,204 |
| 14 | pyspur | 5,626 |
| 15 | genkit | 5,218 |
| 16 | ragapp | 4,373 |
| 17 | voltagent | 4,207 |
| 18 | agentheroes | 3,383 |
| 19 | agents | 2,928 |
| 20 | genaiscript | 2,859 |
| 21 | lmnr | 2,478 |
| 22 | langgraphjs | 2,362 |
| 23 | ruler | 2,221 |
Sponsored
SurveyJS: Build JSON-Driven Surveys and Forms with Full Data Control
Add the SurveyJS UI components to your JS app (React/Angular/Vue3). Securely collect and analyze data without sending it to 3rd-party servers. Fully customizable, works with any backend, and ideal for data-heavy apps. Learn more.
surveyjs.io