Skip to content

docs: clarify stateful server limitations#254

Open
haosenwang1018 wants to merge 1 commit intophilschmid:mainfrom
haosenwang1018:docs/stateful-server-limitations
Open

docs: clarify stateful server limitations#254
haosenwang1018 wants to merge 1 commit intophilschmid:mainfrom
haosenwang1018:docs/stateful-server-limitations

Conversation

@haosenwang1018
Copy link
Copy Markdown

Summary

  • document that mcp-cli is request-oriented, not a persistent interactive session runner
  • explain why browser/session-heavy MCP servers are a poor fit for separate mcp-cli call ... invocations
  • point users toward persistent MCP clients/runtimes for durable browser state

Problem

Issue #21 reports that stateful servers such as chrome-devtools-mcp are hard to use through repeated mcp-cli call ... commands because browser/session state is not guaranteed to behave like one long interactive session.

Changes

  • add a README note under connection pooling that the daemon is an optimization, not a persistent interactive session model
  • add a dedicated Stateful server limitations section explaining which server types are affected and why
  • recommend native/persistent MCP clients for browser/session workflows that need durable state across steps

Fixes #21

Validation

  • git diff --check

Signed-off-by: haosenwang1018 <haosenwang1018@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Stateful servers like chrome-devtools don't work with mcp-cli

1 participant