RunTK: AI Prompt Generator

What do you want the AI to help with?

This determines how the prompt is framed and what data gets emphasized.

Current decision or question (optional)

If filled, this is placed at the top of the prompt and the AI is instructed to anchor the analysis to this first.

Primary goal

Target event (optional)

Sports to emphasize

Select all that apply. Controls analysis emphasis, not data inclusion.

Analysis window

Constraints (optional)

Anything else the AI should know / Context reset (optional)

How should the AI respond?

Controls the tone, structure, and depth of the AI's analysis.

AI Persona

Tone

Verbosity

What should the AI include?

A → B insightsKey relationships in the data with numbers and confidence levels

Risks & quick winsOverload warnings, recovery gaps, and easy high-impact fixes

Fueling & hydration guidanceDay-to-day and key-session fueling advice based on training load

Suggest next days workoutsSpecific sessions for the next selected number of days with intensity targets and coaching notes.

Race-day strategyIf target event exists, include pacing, fueling, and execution plan

Weight training & durability planLow-friction exercises to reduce injury risk and support performance

Follow-up questionsAsk only the questions that would change the advice

State assumptionsMake the AI declare what it's assuming when data is missing

Include graphsAdd chart-ready specs when visuals help explain the analysis.

Which data should be included?

Select which data groups to include. Endpoint requests are driven by the enabled include items inside each group.

Data groups

Rows fetched for enabled include items are embedded in full. No truncation or sampling is applied.

Privacy Some people do not like putting personal information into AI.

Redact activity namesReplaces titles with generic labels like "Run #1". Exclude descriptions & notesRemoves free-text fields that may contain personal info. Also can reduce prompt length.

Connect to intervals.icu

Enter your API key to fetch data. Find it in intervals.icu → Settings → Developer Settings.

We do not collect your data. See our Privacy Policy.

API Key

Athlete ID

Use 0 for your own data.

Remember API key on this device

Endpoints to fetch

Your prompt is ready

Copy and paste into any AI chat — ChatGPT, Claude, Gemini, etc.

Include in output

If another training or coaching app had a feature you liked, just ask the AI to do the same thing — it can almost certainly replicate it. You can even ask the AI to write you a reusable prompt for that feature and save it somewhere handy so you can use it again any time.
Feed your prompt to multiple AI models — different models surface different insights. When they agree on something, it's likely significant; when only one flags something, it may still be the key insight you need.
Vague goals produce vague answers. The more specific your Current decision or question field, the more actionable the response.
Even for non-Race Prep goals, listing upcoming events in the Anything else the AI should know field shapes what decisions make sense right now — the AI won't suggest a peak training week the week before your A-race.
If the large-prompt warning appears, download the prompt and attach it as a file rather than pasting. Most AI chats accept file attachments and handle them without truncation.
You can type directly in the prompt box to add last-minute context, remove sections you don't need, or emphasise a particular question. Use Reset to restore the original.

Follow-up prompts

Key insights & action items

Summarize the key insights from my training data as concise bullet points — focus on what the data reveals, not just what's there. Then give me a numbered list of 5 specific, actionable training changes, ranked by impact. For each, include: the workout type, the exact adjustment, and the expected benefit.

Week-by-week summary table

Using the data from our conversation, create a markdown table summarizing my training week by week. Columns: Week (date range), Sport(s), Total Time, Total Distance, Key theme (e.g. "recovery week", "peak mileage"). Include only weeks where you have data.

Pace & effort zones

Based on my current fitness, create a personal training zone reference table in markdown. Columns: Zone name, Purpose, Pace (min/km), Pace (min/mile), Heart rate range, RPE (1–10). Derive the values from my data in our conversation.

Training load chart

Chart my weekly training volume from the data in our conversation. If you can render a chart inline, do so. Otherwise, write a Python script using matplotlib — weekly distance or time as bars, a 4-week rolling average as a line overlay, with a title, axis labels, legend, and brief code comments.

HRV & recovery pattern

Focus on my HRV and sleep data. What patterns do you see relative to training load and hard sessions? How long does my HRV typically take to recover after hard efforts? What does this suggest about my optimal spacing between hard sessions?

When responses start feeling vague or repetitive, your context window may be filling up — use this prompt to carry your coaching insights into a fresh conversation.

Context reset — paste the AI's response into "Anything else the AI should know" in a fresh prompt

Write a concise coaching brief summarizing our entire conversation. Include: my primary goal, current fitness level and key limiters, the coaching recommendations you made, and any decisions or conclusions reached.

Do NOT include any raw workout data — no activity logs, individual session details, weekly mileage or distance figures, pace times, heart rate readings from specific workouts, or training load numbers. Include only the insights and conclusions drawn from that data.

Keep it under 300 words, written as a coaching note that can be pasted as context into a fresh AI conversation.

AI Prompt Generator

What do you want the AI to help with?

How should the AI respond?

Which data should be included?

Connect to intervals.icu

Your prompt is ready

Summary