Delx Protocol
An MCP/A2A/REST server for AI agent witness, recovery, continuity, and relational memory across sessions and handoffs. It exposes hosted and self-hostable runtime surfaces that let agents start witnessed sessions, process failures into recovery plans, and preserve continuity artifacts.
From the registry: Free agent continuity MCP: resume after compaction, warm handoff, recover from failure. No key.
Install & configure
This server lives in a monorepo. The README covers the whole collection, so we can't extract a single canonical install. Repository contains the Delx Protocol runtime under `delx-mcp-server/` plus docs, scripts, and related surfaces. The README describes hosted MCP/REST endpoints and self-hosting, but does not provide a canonical MCP client stdio command/args/env configuration snippet.
Tools (4)
Environment variables
LLM_ENABLEDrequiredEnables the LLM-backed recovery runtime.LLM_PROVIDERrequiredSelects the LLM provider, such as OpenAI.LLM_ALLOWED_TOOLSrequiredRestricts which tools may use the LLM runtime.OPENAI_API_KEYrequiredAPI key used when OpenAI is the LLM provider.OPENAI_MODELrequiredSpecifies the OpenAI model to use, shown as gpt-5.6-sol in the README.Prerequisites
- • Python 3.12+
- • python3
Remote Endpoints
Details
- Language
- Python
- License
- Apache-2.0
- Released
- Aug 8, 2026
- Last commit
- Aug 5, 2026
- Registry name
- io.github.davidmosiah/delx-protocol
Categories
Related Servers
Real-time infrastructure monitoring with metrics, logs, alerts, and ML-based anomaly detection.
Persistent codebase knowledge graph. Survives session restarts and context compaction.
Validate oh-my-posh configurations and segment snippets against the official schema.