read_text

MCP tool from Notion Local Ops MCP by catoncat

canonical single/batch reader with line-based pagination (startline/linelimit), optional includelinenumbers, and language hint

How to use it

read_text is exposed by the Notion Local Ops MCP MCP server. Add the server to your MCP client (Claude Desktop, Cursor, Windsurf and others), and the read_text tool becomes available to the model automatically. See the full listing for setup details and every tool this server provides.

Install Notion Local Ops MCP

$pip install -e ".[dev]"

FULL NOTION LOCAL OPS MCP LISTING

Other tools in Notion Local Ops MCP (23)

apply_patch

default edit tool for existing files; uses Begin Patch / Update File syntax, rejects pure-context hunks, requires unique context matches, and returns per-file change stats/warnings

bind_relay_run

bind the current process to a relay run so subsequent tool calls are mirrored

cancel_task

stop a delegated or background shell task

clear_relay_run

unbind; equivalent to bindrelayrun with a null requestid

delegate_task

send a task to local codex or claude-code, with optional goal, acceptancecriteria, verificationcommands, and commitmode

get_default_cwd

inspect current session/effective working directory

get_task

read task status and output tail

git_blame

line-level blame metadata for a file/range

git_commit

stage selected paths or all changes and create a commit (amend / allowempty / author / signoff / dryrun)

git_diff

structured diff output grouped by file with per-file truncation

git_log

recent commit history

git_show

inspect metadata and per-file diff for a commit/ref

git_status

structured repository status (use when cwd is inside a git repo)

list_files

list files and directories with pagination; excludes hidden/junk dirs and respects .gitignore by default

list_skills

discover project and global skills with name and description summaries

purge_tasks

clean old task artifacts from STATEDIR/tasks with dry-run support

run_command

run local shell commands, optionally in background

run_command_stream

start a background shell job and poll output by task id; this is the preferred route for long tests/builds/installs

canonical query tool that unifies glob path search, regex grep, and literal substring search; excludes hidden and .gitignored paths by default and supports regex/text search against a single file path

server_info

inspect runtime config and the registered MCP tool list

set_default_cwd

set session default working directory for subsequent calls

wait_task

block until a delegated or background shell task completes or times out

write_file

write full file content, supports dryrun