Boardsesh Climbing Logbook

io.github.wmmg101Published by GitHub account wmmg101v0.1.0Updated Oct 24, 2026

Read-only access to your Boardsesh logbook across Kilter, Tension, MoonBoard and more

VerifiedSTDIODesktop onlyLocation & LifestyleData & Analytics

Overview

AI-generated overview

An unofficial, read-only MCP server that lets an assistant read and analyse your Boardsesh climbing logbook across Kilter, Tension, MoonBoard and other boards.

What you can do
Summarise totals, ascents, sessions and projects across all your boards, filterable by board, angle, outcome and date range. Compare boards with normalised grades, and show grade pyramids, flash rates and month-by-month progression. With optional login: recommend climbs, search the catalogue, find similar climbs and view a hold heatmap.
When to use it
Useful if you log climbs on Boardsesh and want an assistant to answer questions about your sessions, projects, grade pyramid, progress over time, or which board you climb stronger on. Best when you climb on more than one board, since it can compare them.
Requirements
A local Python runtime with uv installed, run via uvx. The required BOARDSESH_USER environment variable (your public display name or user id). Optional BOARDSESH_EMAIL and BOARDSESH_PASSWORD unlock four recommendation, search and heatmap tools; optional BOARDSESH_TIMEZONE sets day grouping. Network access to Boardsesh.
Before you install
Read-only; nothing is written to your account. The default setup needs no secret, only your public display name. If you add BOARDSESH_EMAIL and BOARDSESH_PASSWORD, they are used once to obtain an in-memory token; credentials and tokens are not written to disk or shown in output. Data goes only to Boardsesh endpoints. Unofficial project, not a Boardsesh product.

Installation

In SourceWeft

  1. Open Boardsesh Climbing Logbook in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

boardsesh-mcp

An unofficial MCP server that lets your AI agent read and analyse your Boardsesh climbing logbook — across every board you climb on: Kilter, Tension, MoonBoard, Decoy, Touchstone, So iLL, Woods and spray walls.

How did my last session go? Am I stronger on Kilter or on Tension? What are my projects at 40°? What should I try next at my level? Show my grade pyramid. Which holds do I avoid?

Read-only. Nothing is ever written to your Boardsesh account.

You do not need to give it a password. Boardsesh serves your logbook publicly, so your display name alone unlocks most of it. Credentials are optional and only add the extras.

Add to your agent

Kiro

Open your MCP config (command palette → Kiro: Open user MCP config (JSON)) and add:

json
{  "mcpServers": {    "boardsesh": {      "command": "uvx",      "args": ["boardsesh-mcp"],      "env": {        "BOARDSESH_USER": "Your Boardsesh Display Name"      },      "disabled": false,      "autoApprove": []    }  }}

That is the whole setup. BOARDSESH_USER is the display name on your Boardsesh profile (or your user id, if you prefer). Requires uv installed; uvx ships with it.

Other clients

Same three facts everywhere: command uvx, args ["boardsesh-mcp"], env BOARDSESH_USER. Copy-paste configs for Claude Desktop, Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI and Zed are in docs/clients.md.

Optional: unlock the rest

Four tools need to know which physical board setup you use (layout, size, hold sets), which Boardsesh only tells the logged-in owner. Add your Boardsesh login to enable them:

json
"env": {  "BOARDSESH_USER": "Your Display Name",  "BOARDSESH_EMAIL": "${BOARDSESH_EMAIL}",  "BOARDSESH_PASSWORD": "${BOARDSESH_PASSWORD}"}

Those four are boardsesh_recommend_climbs, boardsesh_search_climbs, boardsesh_find_similar_climbs and boardsesh_get_hold_heatmap. Everything else works without them. See how to set environment variables.

Optional: timezone

Sessions are grouped by calendar day in your timezone, detected from the machine running the server. If that is wrong, add "BOARDSESH_TIMEZONE": "Europe/Rome" (any IANA name).

Example questions

text
Summarise my Boardsesh logbook.How did my last session go?Compare my climbing on Kilter and Tension — where am I actually stronger?What are my current projects, and which have I tried the most?Show my grade pyramid and my flash rate per grade.Have I improved over the last six months?What should I try next at my level?Find climbs similar to the one I sent yesterday.Which holds show up most in climbs one grade above me?

Tools the agent gets

Works with just BOARDSESH_USER:

ToolWhat it returns
boardsesh_get_summaryTotals across all boards: sends, flashes, sessions, date range, hardest send
boardsesh_get_ascentsLogbook entries, filterable by board, angle, outcome and date range
boardsesh_get_sessionsEntries grouped by day
boardsesh_get_projectsClimbs attempted but never sent, per board, climb and angle
boardsesh_compare_boardsOne row per board, with cross-board normalised grades
boardsesh_get_grade_pyramidSends per grade with flash rates
boardsesh_get_progressionMonth-by-month or week-by-week trend
boardsesh_get_gradesA board's grade scale (difficulty ids to labels)

Needs BOARDSESH_EMAIL / BOARDSESH_PASSWORD:

ToolWhat it returns
boardsesh_recommend_climbsClimbs to try next: at your level, crowd favourites, hidden gems, fresh
boardsesh_search_climbsCatalogue search by name, grade range, benchmarks, popularity
boardsesh_find_similar_climbsClimbs sharing holds with a given climb
boardsesh_get_hold_heatmapWhich holds the wall's climbs use, for spotting weaknesses

How the data is interpreted

  • One logbook entry is one climb, at one angle, on one board, at one time, with a number of tries. Entries are never deduplicated, so a climb can appear many times.
  • status is authoritative: flash (sent first try), send (sent after attempts) or attempt (not sent). It is never inferred from try counts.
  • A project is a (board, climb, angle) you have attempted but not sent. Sending it at another angle does not remove it.
  • grade is the board's own grade and difficulty_id the raw id. Board grades are not comparable between boards, so boardsesh_grade — Boardsesh's cross-board normalised number — is what boardsesh_compare_boards uses. It is null when Boardsesh's confidence in it is not backed by real ascents.
  • A "session" is one calendar day in your timezone.

Security and privacy

  • In the default setup there is no secret to leak: only your public display name is configured, and everything it reads is what your Boardsesh profile already shows publicly.
  • If you do add credentials, they are used once to obtain a token from Boardsesh's own login endpoint. Tokens live in memory for the life of the process and are never written to disk.
  • Credentials and tokens never appear in tool output, logs or error messages.
  • The server talks only to ws.boardsesh.com and www.boardsesh.com. No telemetry.
  • Every tool is read-only, and the GraphQL layer can only send a fixed set of pinned read queries — there is no code path that could reach a mutation.
  • Your logbook is fetched at most once a minute however many tools the agent calls.

Troubleshooting: run uvx boardsesh-mcp --check with the same environment variables set. It prints counts and configuration only, no entries and no secrets, so it is safe to paste into an issue.

Relationship to Boardsesh and to kilter-mcp

Boardsesh is an open-source (Apache-2.0) board-climbing app that aggregates logbooks across boards and publishes an API. This project is an independent MCP client for it, built at the suggestion of its maintainer, but it is not an official Boardsesh product and any bug here is mine, not theirs.

If you only climb on a Kilter Board and do not use Boardsesh, kilter-mcp talks to Kilter directly instead. boardsesh-mcp is the better choice if you use more than one board, because it can compare them.

Contributing

Issues and pull requests welcome — see CONTRIBUTING.md. Security reports go through SECURITY.md.

License

MIT

Source: README.md at commit ae71813

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v0.1.0LatestSep 22, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts