Agent Handoff Protocol (AHP)

Product · Developer tool by DeepJudge

Context continuity for the agentic AI ecosystem.

An open-source protocol for AI systems that enables context continuity, allowing users to move between different AI platforms without losing their conversation history, materials, or work progress.

Launched
2026-08-13
Languages
English

Category: Open Source

License
Open protocol
Protocol Type
Open Protocol
Primary Use Case
Context continuity across AI systems
Developer Access
GitHub repository
Initial Partners
Harvey, Thomson Reuters
Release Date
August 13, 2026
  • Open protocol for context continuity across AI systems
  • Preserves conversation history and prior work product during handoffs
  • Eliminates the need to re-upload materials when switching AI tools
  • Supported by Harvey and Thomson Reuters
  • Available as an open standard on GitHub

Overview

Agent Handoff Protocol (AHP) is an open-source protocol developed by DeepJudge to enable context continuity across diverse AI systems. It is designed for AI providers, specialist platforms, and law firms building in-house AI solutions to ensure that users can move between specialized products without losing their work thread. The protocol addresses the "context tax" by transferring objectives, supporting materials, conversation history, and prior work product from one system to another. For instance, a lawyer can transition from researching a precedent in DeepJudge to drafting in Harvey with all relevant context already in place. Released in August 2026, AHP is supported by industry leaders including Harvey and Thomson Reuters to foster an interoperable, agent-driven legal technology ecosystem.

Context Management

  • History Preservation: It preserves the conversation history and prior work product that led to the current point when moving between AI experiences.
  • Material Transfer: The protocol carries supporting materials and objectives across systems, eliminating the need to re-upload documents.
  • Seamless Transitions: AHP allows users to move between specialized AI products without losing the thread of their work or starting fresh.

Developer and Ecosystem Features

  • Open Standard: DeepJudge developed AHP as an open protocol to help advance interoperable, agent-driven legal technology.
  • GitHub Availability: The protocol specification and implementation resources are publicly available on GitHub for adoption.
  • Protocol Interoperability: AHP is designed to work alongside existing protocols like the Model Context Protocol (MCP) to provide broader task context.
  • Shared Resource Model: The protocol is offered as a shared resource rather than a proprietary advantage to invite industry-wide adoption.

Use cases

  • Lawyers move from researching a precedent in DeepJudge to drafting a document in Harvey with all context preserved.
  • Developers implement the protocol to allow their custom AI agents to exchange information with third-party legal AI platforms.
  • Legal teams switch between specialized AI tools for different tasks without re-uploading documents or recreating prompts.

Specifications

Integrations
Harvey, Thomson Reuters, DeepJudge Knowledge Search, DeepJudge AI Workflows
License
Open protocol
Protocol Type
Open Protocol
Primary Use Case
Context continuity across AI systems
Developer Access
GitHub repository
Initial Partners
Harvey, Thomson Reuters
Release Date
August 13, 2026

How it works

  1. Access Specification: Developers access the open protocol specification and implementation resources via the official GitHub repository.
  2. Implement Protocol: AI providers or law firms building in-house solutions implement the protocol to enable their systems to send or receive context packets.
  3. Execute Handoff: Once integrated, a user can initiate a handoff from one AI tool (like DeepJudge) to another (like Harvey) while maintaining their full task context.

FAQ

What is the Agent Handoff Protocol?

The Agent Handoff Protocol (AHP) is an open protocol designed to enable context continuity across AI systems, allowing users to move between platforms without losing their work progress.

What specific information does AHP transfer?

AHP enables the transfer of user objectives, supporting materials, conversation history, and prior work product from one AI system to another.

Is the Agent Handoff Protocol open source?

Yes, DeepJudge has released AHP as an open protocol, and the specification and implementation resources are available on GitHub.

Which platforms support the Agent Handoff Protocol?

Harvey and Thomson Reuters are among the first to adopt the protocol, with Harvey's integration entering beta in August 2026.

Added by AntSpace from public sources (deepjudge.ai).

More products by DeepJudge