claude-web-history-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| CLAUDE_ORG_ID | No | Skip org lookup by pinning your org uuid. | |
| CLAUDE_WEB_UA | No | Override the User-Agent string. | |
| CLAUDE_WEB_BASE | No | Override the base URL. Must still resolve to claude.ai. | |
| CLAUDE_SESSION_KEY | No | The claude.ai session key. If not set, the server uses the OS keychain. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_web_conversationsA | List conversations from the claude.ai web account (newest first). Returns uuid, title and timestamps. Use read_web_conversation to pull one in. |
| read_web_conversationA | Fetch the full transcript of one claude.ai conversation as markdown, so it can be used as context in this Claude Code session. |
| save_web_conversationA | Fetch a claude.ai conversation and write it to disk as markdown — useful for parking it in the repo (e.g. docs/context/) so it persists across sessions. |
| debug_connectionA | Check whether the session key works and which org id is resolved. Optionally hit a raw API path to inspect its shape when an endpoint changes. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 4 tools
The tools are mostly distinct: debug_connection checks credentials, list_web_conversations lists metadata, read_web_conversation fetches content, and save_web_conversation persists to disk. The only potential overlap is read vs save, but the descriptions clarify the output destination (context vs file).
All tool names follow a consistent verb_noun pattern in snake_case: debug_connection, list_web_conversations, read_web_conversation, save_web_conversation. The singular/plural variation is natural and doesn't create confusion.
With only 4 tools, the set is tightly scoped to the server's purpose of accessing and persisting claude.ai web conversations. Each tool serves a clear, non-redundant function, and the small count is appropriate for the niche domain.
The core lifecycle is covered: list conversations, read a conversation, and save a conversation. Debugging is supported via debug_connection. The only potential gaps are search or deletion, but those are arguably outside the intended scope of a history-retrieval tool.