Claude Code
Best for: Serious work on real repos
Anthropic's agent for your terminal, IDE and desktop. Reads the whole repo, runs commands and tests, and handles large multi-file changes well.
Visit Claude Code ↗Home / Use case
Give your AI agent a memory that lasts
“My agent forgets everything between sessions and I keep re-explaining my project.”
A plain notes file the agent reads at the start and updates at the end beats most fancy memory systems. Keep it short, or it overflows and the agent gets confused.
Fix it yourself in 3 steps
- Create one notes file in your project (decisions, how to run things, open to-dos) and tell the agent to read it first every session.
- At the end of each session, ask the agent to propose edits to that file; you approve what gets saved.
- Once a month, have it trim stale or duplicate notes so the file stays small.
Copy this prompt
Read NOTES.md before doing anything. When we finish, propose changes to NOTES.md: new decisions, commands that worked, and open to-dos, under 40 lines total. Show me the diff and wait for my OK before saving. Never store passwords or keys in it. Stop after showing the diff.
Fill in the [brackets]. Each prompt says when the AI should stop, so it doesn't wander off.
Agents people used for this: Claude Code, ChatGPT.
Inspired by real examples collected at agentusecases.ai.
📚 4 real stories of people doing this with AI agents
- Add agentic forgetting to an agent's memory · custom-built or unnamed agent
- Give an agent persistent memory and identity · Claude Code
- Give coding agents git-native persistent memory · custom-built or unnamed agent
- Save AI Q&A summaries into git repos as persistent notes · ChatGPT
Collected at AgentUseCases.ai.
Tools for this
ChatGPT
Best for: Widest feature set
OpenAI's assistant with the broadest feature set: voice, images, search and a large app ecosystem.
Visit ChatGPT ↗