CONTINUUM: Verifiable semantic recovery for long-running AI agents. Semantic checkpoints (not conversation dumps), an idempotent action ledger that refuses duplicate side effects, and a hash-chained tamper-evident event log, all exposed as a deny-by-default MCP server. Framework-agnostic, Python 3.11+.
View on GitHub →
🔍 Audit Wallet Slippage →
Quick Install
Copy the config for your editor. Some servers may need additional setup — check the README.
Claude Desktop
Claude Code
Cursor
Add to claude_desktop_config.json:
{
"mcpServers": {
"continuum": {
"command": "uvx",
"args": [
"continuum"
]
}
}
}📋 Copy
Run in terminal:
claude mcp add continuum uvx continuum📋 Copy
Add to .cursor/mcp.json:
{
"mcpServers": {
"continuum": {
"command": "uvx",
"args": [
"continuum"
]
}
}
}📋 Copy
Or install with pip: pip install continuum
README Excerpt
<p align="center"> <img src="docs/assets/readme-img.png" alt="CONTINUUM Banner" width="100%" /> </p> <p align="center"> <strong>CONTINUUM: Verifiable semantic recovery for long-running AI agents.</strong> Semantic checkpoints (not conversation dumps), an idempotent action ledger that refuses duplicate side effects, and a hash-chained tamper-evident event
Tools (15)
BrowserAdapter ContainerAdapter FilesystemSandboxAdapter GenericAgentAdapter KubernetesAdapter LangChainAgentAdapter LangGraphAgentAdapter OpenAIAgentAdapter PythonInProcAdapter action_index checkpoints events events_archive runs versions
Topics
agent-framework ai-agents checkpointing crash-recovery durable-execution fault-tolerance hacktoberfest langchain langgraph llm llm-agent mcp mcp-server model-context-protocol observability