AltGrowthDiagnose · Plan · Execute
A growth operator that actually does the work.
Paste your URL. AltGrowth reads your site, opens with specific hypotheses about what's capping growth, generates a critiqued 90-day plan, and runs the weekly cadence with you — learning from every experiment.
Free. About five minutes. URL optional, but worth it.
The loop
Three rungs. One operator. Compounding learning.
Diagnose
Streaming research + a planner-aware chat that asks only what it needs.
Plan
A critiqued 90-day plan: objectives, hypotheses, weekly experiments, KPIs.
Execute
Track experiments week by week — the agent learns from each result.
Diagnose
It reads your site before you say a word.
Most tools start with a 20-question intake. AltGrowth scrapes your site, infers your stack and ICP, and opens with a hypothesis — then asks pointed questions to confirm or kill it.
Pre-intake research
Site scrape, tech-stack inference, and signal gathering kick off the moment you paste a URL — so the first question is already informed.
Generative diagnostic cards
ICP, positioning, funnel-leak, and channel-fit cards render as the AI thinks — not a wall of text. Validated with Zod schemas.
Planner-aware agent
The chat tracks subtopic state and only asks what's missing. No 'let's go through 12 questions' — it skips ahead when it already has the signal.
Synthesis reframe
At the end of intake, the agent presents a one-screen reframe of your real growth bottleneck — your chance to push back before a plan is generated.
Plan
A 90-day plan that survived its own peer review.
Every plan goes through a three-stage critique pipeline before it ever lands in your dashboard. Then it's yours to edit, version, and steelman.
01
Critique findings
The diagnostic gets stress-tested before any plan is written. Weak signals get flagged.
02
Generate (strict JSON)
Objectives, hypotheses, experiments, KPIs, and weekly cadence — schema-validated end-to-end.
03
Audit & score
A second-pass judge scores the plan across dimensions. Below threshold? It regenerates.
Per-experiment regenerate
Don't like one experiment? Swap it. The agent regenerates just that card with your constraint, keeping the rest of the plan intact.
Steelman any objective
Ask the agent to argue the strongest version of an objective — before you commit a quarter to it.
Versioned, reversible
Every regenerate, edit, and reorder writes an append-only snapshot. Restore any prior version with one click.
Reorder & inline edit
Drag to reorder. Edit titles, hypotheses, and KPIs in place. The plan is a working document, not a PDF.
Execute
Where most growth tools quit, AltGrowth shows up every Monday.
A plan is worthless if it sits in a Notion doc. AltGrowth tracks status per experiment, prompts a weekly check-in, and feeds outcomes back into the agent so the next regenerate is smarter than the last.
Status per experiment
Not started · running · shipped · dropped. Dashboard rollup chips show the state of every plan at a glance.
Weekly check-ins
A lightweight modal asks what moved, what stalled, and what you learned. Two minutes a week.
Results drawer
Capture wins, losses, and surprise findings against the original hypothesis. Outcomes become structured data.
Monday-morning nudge
If you have a running experiment, you get a per-plan email at 9am local — week N of 12, day counts, one-click check-in.
The adaptive loop
The agent that ran your last experiment writes your next one.
Adaptive regenerate
When you regenerate an experiment after running a few, AltGrowth pulls in your prior wins, losses, and weekly check-ins as context. The version history shows a 3r·5c chip so you know exactly which runs informed the rewrite.
Past runs
folded in
Check-ins
weighted
Reasoning
on the chip
Pricing
Diagnose for free. Pay only when you're ready to ship.
Diagnose your business and get a critiqued 90-day plan. Keep one plan in your dashboard.
- · Streaming research + diagnosis chat
- · Critiqued 90-day plan, exportable
- · Inline editing, drag-to-reorder, version history
Run the weekly loop. Adaptive regenerate that learns from your wins and losses. Monday-morning nudges. Renew when you're ready — never auto-charged.
- ✓ Mark experiments running / shipped / dropped
- ✓ Weekly check-ins & result drawers
- ✓ Steel-man & per-experiment regenerate
- ✓ Adaptive regenerate informed by past runs
- ✓ Monday email nudge with active experiments
Stop guessing. Start the loop.
Five minutes from URL to a 90-day plan, free. Then $199 / 90 days one-time when you’re ready to run the weekly Execute loop. No subscription, no auto-charge.
Free. About five minutes. URL optional, but worth it.