thred

$20/moStarting price

0Popularity

About thred

thred provides a shared memory system for AI agents such as Claude, Codex, and Cursor. It enables agents to save and resume work by preserving decisions, evidence, and next steps in a continuous context stream. Each agent connects through a workspace that maintains state, allowing subsequent agents to pick up where previous ones left off without losing context. The system uses MCP (Model Context Protocol) to integrate with existing agent workflows, ensuring compatibility with tools like Windsurf, Cline, and others. Users configure MCP connections and agent instructions to establish shared memory across different agents and tasks. The workspace tracks changes and resolves context updates automatically, reducing the need for manual handoffs or repeated instructions.

Key features

Shared memory stream for agents
MCP-based integration
Automatic context resolution
Agent instruction configuration
Workspace state tracking
Checkpoint resumption
Cross-agent handoffs
Decision and evidence preservation

Use cases

Continuous project development across multiple AI agents
Seamless handoff between different coding or writing tools
Maintaining context in long-running agent workflows

Pros

Preserves agent context across sessions
Supports multiple AI agents in a single workspace
Integrates via MCP for compatibility with existing tools
Tracks decisions and next steps automatically
Reduces manual handoffs between agents

Cons

Requires MCP configuration for use
Limited to agents supporting MCP integration

Frequently asked questions about thred

What is thred and what does it do?

thred provides a shared memory system for AI agents such as Claude, Codex, and Cursor, enabling them to save and resume work by preserving decisions, evidence, and next steps in a continuous context stream.

How does thred work with different AI agents?

Each agent connects through a workspace that maintains state, allowing subsequent agents to pick up where previous ones left off without losing context. The system uses MCP (Model Context Protocol) to integrate with existing agent workflows.

Which AI agents or tools are compatible with thred?

thred is compatible with agents and tools such as Claude, Codex, Cursor, Windsurf, and Cline through MCP integration.

How do I set up thred for my agents?

Users configure MCP connections and agent instructions to establish shared memory across different agents and tasks. The workspace tracks changes and resolves context updates automatically.

Can thred preserve the state of my work across different sessions?

Yes, thred preserves the state of work in motion, including decisions, evidence, and next steps, so the next agent can resume where the previous one stopped.

What is MCP and how does it relate to thred?

MCP (Model Context Protocol) is a protocol used by thred to integrate shared memory into the tools and agents you already use, ensuring compatibility and seamless handoffs.