convex-agents

Build persistent, stateful AI agents with Convex and OpenAI.

Updated Feb 26, 2026

Search Tags:#ai #rag #agents #streaming #convex #tool-integration

Ready to run

npx skills add https://github.com/NitzanSelwyn/MomentumTrackerAdmin --skill convex-agents-nitzanselwyn

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: convex-agents
Source: https://github.com/NitzanSelwyn/MomentumTrackerAdmin/tree/main/.claude/skills/Convex%20Agents
Command: npx skills add https://github.com/NitzanSelwyn/MomentumTrackerAdmin --skill convex-agents-nitzanselwyn

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Build persistent, stateful AI agents that maintain context across sessions and enable long-running workflows.

Core Features & Use Cases

Thread management for multi-turn conversations with persistent state
Tool integration to perform actions via Convex functions
Streaming responses to clients for real-time UX
RAG patterns for knowledge retrieval and contextual responses
Workflow orchestration for complex, multi-step tasks

Quick Start

Install the Convex agent package, configure your OpenAI API key, and instantiate the Agent to start building a chat-enabled AI workflow.

Frequently Asked Questions about convex-agents

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I build persistent AI agents that maintain state across live chat sessions?▼

Persistent AI agents are built using the Convex Agent component to manage thread state, integrate tools, and orchestrate workflows across live chat sessions. This enables context retention and long-running conversational workflows.

How do I stream AI responses to clients in real time during a chat session?▼

Streaming responses to clients is handled natively by the Convex Agent toolkit, pushing real-time AI outputs directly to the frontend during live chat sessions for an interactive user experience.

What's the best way to implement RAG patterns for AI agents in a Convex project?▼

Implementing RAG patterns in a Convex project involves using the Convex Agent component for knowledge retrieval and contextual responses. This integrates directly with your existing stateful agent workflows.

Do I need an OpenAI API key to use Convex agent components?▼

An OpenAI API key or access to compatible LLMs is required to run tools, stream outputs, and persist conversations. You also need the @convex-dev/agent toolkit installed in a Convex-enabled project.

Can I orchestrate multi-step workflows and tool integration using AI agents in Convex?▼

Workflow orchestration and tool integration are supported natively in Convex AI agents. You can execute complex multi-step tasks by triggering Convex functions as tools within your agent workflows.