If you are working by yourself, no need to over complicate it, I have three terminal sessions open - Claude, Codex and one for testing that has my AWS credentials.

I tell both Claude Code and Codex to keep the same markdown file updated where I keep my progress, requirements, decisions, thought process etc updated

undefined

Anthropic is pushing MCP for a reason. The idea is to spin up a local MCP server that serves project context - docs, architectural decisions, DB schemas - which Claude Desktop and Cursor (via plugins or natively soon) can connect to

For a team this is a total killer feature: you spin up a shared MCP server with documentation, and all agents instantly sync their "brains"

for what its worth - i've settled on a simple markdown file that i keep updated with project context, decisions, and quirks. before starting any AI session i just paste the relevant sections in. not elegant but it works.

some folks have been experimenting with MCP (model context protocol) servers that sync state between tools. haven't tried it personally but the idea is you have a single source of truth that all your agents can query.

I use markdown files which full detail and summarized versions. I also create files just for an aspect of it (security, performance) so that it makes it easier to reference.

I saw before that someone has developed a software that can transform the context between major AI programs, which should be helpful to your problem.