Show HN: Drop-In Checkpointer for LangGraph (No DB Setup Required)

3 months ago 2

🧠 Convo SDK — LangGraph Checkpointer

LLM tools forget everything between runs.

Convo SDK is a plug-and-play LangGraph checkpointer for persistent memory storage using PostgreSQL — optimized for AI agents, chatbots, and stateful LLM workflows. ✨ Features 🔌 Seamless LangGraph Integration 🔍 Full Message Logging: Human inputs, AI responses, Tool Calls 🧠 Token Usage Tracking & Analytics 🪄 Flexible Thread Management:

Initialize with thread ID Pass thread ID at runtime for checkpointing ⚡ Get your API Key from the Dashboard

📦 Installation npm install convo-sdk 🚀 Quick Usage 1️⃣ Initialize SDK

import { Convo } from "convo-sdk"; const convo = new Convo(); await convo.init({ apiKey: "your-convo-api-key" });

2️⃣ Create or Reuse Thread

const thread_id = await convo.newThread(); // Or use an existing thread ID. // thread_id is just a unique string

3️⃣ Get Checkpointer

const checkpointer = convo.checkpointer(); // optionally you can pass the thread_id here // e.g. const checkpointer = convo.checkpointer({ thread_id });

4️⃣ Use in LangGraph Workflow

const graph = workflow.compile({ checkpointer });

5️⃣ Run Graph with Thread ID (Runtime)

await graph.invoke(state, { configurable: { thread_id, }, });

6️⃣ Restore from Specific Checkpoint

await graph.invoke(state, { configurable: { thread_id, checkpoint_id: "your-checkpoint-id", }, });

📖 API Overview

Method Description
await convo.init({ apiKey }) Initialize SDK
await convo.newThread() Create a new conversation thread
convo.checkpointer({ thread_id: optional }) Get LangGraph-compatible checkpointer
graph.invoke(state, { configurable }) Run graph with checkpointing + config
convo.getThreadState(thread_id) Get full thread state
convo.listThreads() List all available threads
🎯 Key Points:
You can either:
Provide thread_id at checkpointer creation (persistent across runs),
or
Pass it dynamically during graph invocation (flexible per-run control).
Works seamlessly with LangGraph + LangChain agents.
Supports restoring from historical checkpoints.
💡 Note:
With just one import, Convo automatically tracks and persists every message between users and agents — including inter-agent reasoning steps — so you can:
  • Replay conversations & debugging sessions
  • Visualize agent workflows & reasoning
  • Build richer, stateful AI apps without extra plumbing
Read Entire Article