SECTION 11

On-demand sync

From the maintained repository documentation.

On this page

On-Demand Sync

Trigger a fresh Intervals.icu sync from your phone or browser, then download the data files. No local Python, no scheduled runs. You sync when you want to.

Want automatic sync on a schedule? See json-auto-sync instead.

Running an agentic platform locally? See json-local-sync. No GitHub needed.

Overview

Uses the same GitHub Actions workflow as auto-sync, but without the cron schedule. You tap a link in your repo's README to trigger a sync, and download the resulting files as a ZIP artifact from the completed run.

Flow: Tap Sync Now → Run workflow → Download artifact ZIP → Attach to AI chat.

GitHub connector users: If your AI platform has a GitHub connector, you can skip the download, but refresh, sync, or re-import after the workflow commits, since not all connectors pick up changes automatically. See the connector table.

Prerequisites

Step 1: Get Your Intervals.icu Credentials

  • Athlete ID: Intervals.icu → Settings → bottom of page (e.g., i123456)
  • API Key: Intervals.icu → Settings → Developer Settings → API Key

Step 2: Set Up GitHub Repository

  1. Go to github.com/new
  2. Name it something like training-data
  3. Set to Private (recommended)
  4. Check Add a README file
  5. Click Create repository

Then add these files:

File Location Source
sync.py Root examples/sync.py
auto-sync.yml .github/workflows/ examples/json-auto-sync/auto-sync.yml
README.md Root (replace default) DATA_REPO_README_TEMPLATE.md

Important: Edit auto-sync.yml after copying. Remove the schedule block so only workflow_dispatch remains:

on:
  workflow_dispatch:  # Allow manual trigger

This prevents automatic runs. The workflow only runs when you trigger it.

To create the workflow folder:

  1. Click "Add file" → "Create new file"
  2. Name it: .github/workflows/auto-sync.yml
  3. Paste the workflow content, remove the schedule block
  4. Commit

Step 3: Add Repository Secrets

  1. Go to your repo → Settings → Secrets and variables → Actions
  2. Click New repository secret
  3. Add these secrets:
Secret Name Value
ATHLETE_ID Your Intervals.icu athlete ID (e.g., i123456)
INTERVALS_KEY Your Intervals.icu API key

Optional secrets: see auto-sync SETUP for WEEK_START and ZONE_PREFERENCE.

Step 4: Enable Workflow Permissions

  1. In Settings → Actions → General
  2. Scroll to "Workflow permissions"
  3. Select "Read and write permissions"
  4. Click Save

Step 5: Update README Template

Replace YOUR_GITHUB_USER and YOUR_REPO_NAME in README.md with your actual GitHub username and repo name.

Step 6: First Run

  1. Go to your repo → Actions tab
  2. If prompted, enable workflows
  3. Click on "Auto-Sync Intervals.icu Data"
  4. Click Run workflow → Run workflow
  5. Wait 30–60 seconds for the run to complete

After the run completes:

  • Connector users: Refresh, sync, or re-import so your AI picks up the new files. See the connector table.
  • Download users: Click the completed run → scroll to Artifacts → download training-data ZIP.

From now on, just tap 🔄 Sync Now in your repo's README.

Usage

From your phone

Open your repo in a browser → tap 🔄 Sync Now → Run workflow. When the run completes, download the artifact or refresh, sync, or re-import through your connector as required.

With an AI connector

AI platforms with GitHub connectors can read the repo directly, but refresh behavior varies; some need a manual sync or re-import before they see the new files. Check the connector table for your platform, then refresh accordingly. No download needed once the connector has the current files.

Without a connector

Download the training-data artifact ZIP from the completed run, then attach the JSON files to your AI chat.

Security

Downloading the artifact ZIP and attaching the JSON to a chat takes your data outside GitHub authentication. A connector reads the repository directly, so processing and retention there are governed by that AI platform's terms rather than by Section 11. Where the files go from there depends on what you configure and which AI you point them at. This workflow publishes no GitHub Pages site. Private repos remain private. The workflow and required-secret handling are identical to auto-sync apart from the manual trigger. See Privacy & Security for the full policy.

Updating an Existing Repository

When the workflow template changes, follow Updating an Existing Repository in the auto-sync guide, with these differences. Skip its step 2 (Choose a schedule) and step 8 (Watch the next scheduled run). The current auto-sync.yml includes a schedule block, so in its step 4, remove the schedule: block again before you commit and keep workflow_dispatch:, as in Step 2 of this guide; otherwise the workflow starts running on a schedule. Reapply any other changes you still want as usual.

Troubleshooting

See the auto-sync troubleshooting guide. The same workflow runs in both paths.