We've released a closed beta of NASEBANAL Growth Mobile, the mobile client bundling Target, Recorder and Stopwatch that we introduced in a previous post, for both iPhone and Android, and we're recruiting testers ahead of the full release. And separately, we've launched the NASEBANAL MCP Server, which lets AI clients operate NASEBANAL directly. This post covers both.
NASEBANAL Growth Mobile Closed Beta Release
NASEBANAL Growth Mobile brings the three web apps — Target, Recorder, and Stopwatch — into a single mobile app with one account and one login.
- Dashboard: a home screen summarizing your status across all three apps
- Target: edit your hierarchical target tree right from your phone
- Recorder: review data logged per tag as time-series charts and stats
- Stopwatch: log a lap time to Recorder with a single tap
We're currently preparing for a full store release on both iOS and Android, and ahead of that we're looking for beta testers to try the app on real devices and share feedback.
- iPhone: invitation-based beta testing via TestFlight
- Android: invitation-based beta testing via Google Play's closed testing track
For now, joining either program works through an individual invitation via our contact form, rather than a public link. If you'd like to take part, please send a message through the contact form stating that you'd like to join the Growth Mobile beta, along with your device (iPhone or Android) and the email address you'd like the invite sent to. We'll follow up with the enrollment steps. The email address you provide is used only to send that invite.
If you notice anything during testing, we'd appreciate hearing about it through the same contact form — your feedback will directly shape the final release.
Launching the NASEBANAL MCP Server
Before starting NASEBANAL, we built an app — the one whose idea seeded NASEBANAL Growth — and test users told us data entry felt like a hassle. At the time we'd only shipped a web app, so mobile users were accessing that same responsive web app through their browser too. That feedback made us aware we needed a native mobile app, and pushing one step further from there is what led us to LLM-driven data access — the motivation behind this MCP Server.
The new recipe we added to the Cookbook, Control Target and Recorder directly from Claude Desktop via MCP, is a Remote MCP server that connects Claude Desktop or any MCP (Model Context Protocol) client to your own NASEBANAL account (Target and Recorder). It works the same way as the Figma or Trello MCP connectors: add the connector, log in with NASEBANAL (Auth0) in the browser popup, and you're set — no manual token copy-pasting.
For example, step on the scale in the morning and just tell the chat "Log 70.2kg to Recorder" — done. On the weekend, ask "Summarize last week's weight trend in Recorder" and recorder_stats returns the aggregate, ready to hand to someone as a quick report. What used to be opening a form, picking a tag, and typing a number becomes one round-trip of chat.
Once connected, the following tools become available.
Target
| Tool | Does |
|---|---|
target_list | List all your Targets, with nested cells |
target_get | Get one Target by id |
target_create | Create a Target (optionally as a sub-target via parent_id) |
target_update | Update a Target's title/description/memo |
target_delete | Delete a Target |
Recorder
| Tool | Does |
|---|---|
recorder_record_list | List logged data points, filterable by tag/date range |
recorder_record_create | Log a new data point (auto-creates the tag by name) |
recorder_record_update | Update a logged data point |
recorder_tag_list | List tags — use this to find a tag's numeric id for goals |
recorder_goal_list | List Goals |
recorder_goal_create | Create a Goal on an existing tag |
recorder_stats | Summary stats (count/min/max/average/latest) for a tag — the report-generation tool |
Setup is just adding https://mcp.nasebanal.com/mcp under Settings → Connectors → Add custom connector in Claude Desktop. For the full setup walkthrough and current limitations (sharing/collaboration and bulk operations aren't wired up yet, and there's no silent Auth0 token refresh), see the recipe itself. Here's a short video showing it in action.
The video moves fast and the mouse cursor can be hard to follow, so here's a quick recap of what's happening: the Target side has two targets already defined, Health and English, and Health breaks down into BMI, Running Time, and Buzzer (a quick-reaction test). The video first uses Stopwatch to time a Buzzer reaction and logs it to Recorder, then judges pass/fail from that record. It closes by pulling that same Recorder record through MCP from Claude Desktop and evaluating it there.
This MCP server runs as a Remote MCP server on Cloudflare Workers, and under the hood it's just a thin wrapper around @nasebanal/sdk. The SDK itself is a typed client generated from the OpenAPI spec under our Contract-Driven Development pipeline, so adding one more MCP tool is really just registering an already-generated SDK method as a tool definition. The full API surface is already exposed, so we'll keep adding tools as the need comes up.
While we were at it, we also reorganized the Cookbook's categories into Growth, Evolution and Common. MCP Server doesn't belong to any single Growth app — it's a cross-cutting interface onto NASEBANAL as a whole — so it's filed under Common (Evolution is still a Coming Soon placeholder for now).
Closing
A mobile app and an MCP server reach very different places — a phone screen and an AI client — but both are new front doors onto Growth, built on the same OpenAPI spec as their shared contract. We hope you'll join the beta, and give the MCP server a try.
As NASEBANAL Evolution, we're building a platform that supports the continuous evolution of systems centered on API quality assessment, and we want to keep evolving Growth — the side that supports human growth — right alongside it. Looking ahead, we'd also like to explore analyzing video and other media with AI to extract and manage metrics from it.