Runtool Agent

The Runtool Agent (beta) is the AI training partner built into every workspace. It reads the same picture you see — the athlete profile, the calendar, the training history — and helps you plan: drafting weeks, adjusting workouts, and answering training questions with the athlete's actual data behind every answer.

Open it with the header button or ⌘J. The Runtool Agent is available on the Standard, Coach, and Pro plans. It stays off during a free trial and turns on when your subscription starts.

You are in control#

The Runtool Agent proposes; you decide. When it drafts a plan change, the proposed workouts appear on the Calendar as ghost entries for review — approve them and they become real, reject them and they vanish.

Two modes, switchable per conversation and defaulted in Settings → Runtool Agent:

• Preview before applying — every change waits for your approval. The default.
• Apply directly — changes land immediately; you review after the fact.

What it works with#

The Runtool Agent operates on the workspace it's opened in:

• Reads the athlete's details — zones, health metrics, recent weekly volume, and the About description from athlete settings — plus the calendar and completed training, the same data you see in the app.
• Asks when it can't tell. With no recorded history to read a normal week from, it asks you what yours looks like rather than guessing — you're never asked to set up and maintain a training baseline by hand.
• Drafts and edits workouts — single sessions or whole stretches of plan, structured exactly like hand-made workouts.
• Follows established training methodologies when drafting, and explains the why behind its suggestions.
• Answers questions — "how did the last three weeks compare to plan?", "what's behind my load warning?" — from the workspace data.

It sees nothing beyond the workspace, and its proposals follow the same rules as manual edits.

Training framework#

The framework is the coaching methodology the Runtool Agent follows when it drafts workouts. You set it once in Settings → Runtool Agent, and it applies to every athlete you work with — it's your setting, not a per-athlete one.

• Dynamic is the default. The agent blends proven practices and adapts to the athlete in front of it.
• The named frameworks — Daniels, Lydiard, Norwegian, Pfitzinger, Hansons, Canova, Higdon, 80/20, Galloway, MAF, FIRST, Tinman — follow their published structure.

You pick one framework at a time; they describe different approaches to the same goal, so they don't combine.

Volume periodization sits beside the framework in the same settings: the week-over-week volume curve — build and deload cadence — that workout suggestions follow. Linear is the default for most runners.

Ways to ask#

Type freely, pick from suggested starting points when a conversation is empty, or scope the conversation to what you're looking at. Plan a week, rework a block of days, or talk through a race buildup — the Runtool Agent handles the structure while you make the calls.

Usage#

Runtool Agent usage is included in paid plans with a monthly limit. Your current usage shows as a percentage under Settings → Runtool Agent, and resets each billing period.

Standard and Coach share the same monthly allowance. Pro carries 5× that allowance, for runners and coaches who work with the Agent regularly. If you run out before the period resets, the usage meter offers the upgrade.

FAQ#

Can the Runtool Agent change my plan without me noticing?

Not in the default mode. Preview before applying holds every change for approval, and even in Apply directly, every change is visible on the calendar and raises the same plan-change notifications as a manual edit.

Does it replace a coach?

No. It's a planning partner with perfect recall of the athlete's data. Coaches use it to draft faster; self-coached athletes use it to plan with structure. Judgment stays with you.

What happens when I hit the monthly limit?

The Runtool Agent pauses until your next billing period. Your plan, calendar, and data are never affected.