3 Practical Fable Projects to Run Before July 8

Published July 4, 2026

Unlimited Fable goes away on July 8. After that, Anthropic's most capable model moves to pay-per-use, and you'll think twice before every big ask.

That deadline is actually useful. It forces you to stop treating Fable like a slightly smarter chatbot and start using it for work that genuinely needs its depth: long-context reasoning, cross-app synthesis, and one-shot execution on projects you've been putting off.

Here are three prompts to run while it's still free. Each one is designed to compound: the first finds what's worth running on Fable, the second turns that into a real plan, and the third lets you actually build something big.

Why this matters

Fable is Anthropic's newest and most capable model. Right now you can run it without counting every token. After July 8, you pay per use, which means you want a clear filter for what's Fable-worthy vs. what Opus or Sonnet can handle fine.

Prompt 1: Find your Fable-worthy work

Most people waste their last free Fable days on tasks Sonnet could do. This prompt fixes that. It asks Claude to audit everything it knows about you (projects, docs, memory, context files) and rank the work that actually needs Fable's reasoning depth.

When to use it: First. Before you run anything else on Fable.

How to run it: Paste this into Claude with Fable selected. Let it search your projects, read your docs, and pull from memory. Do not ask it to do the work yet. You want the ranked list first.

You're Fable, Anthropic's most capable model. I'm trying to make the most of unlimited Fable before it goes pay-per-use on July 8.

Your job right now is an audit, not execution.

Look through everything you have access to about me:
- Projects and recent conversations
- Any docs, plans, or context files I've shared
- Your memory of me (preferences, goals, recurring themes)
- Connected apps or tools if you can read them (Slack, Google Drive, Notion, etc.)

Based on that, list the top 5 tasks that are genuinely worth running on Fable instead of Opus or Sonnet. For each one, give me:
1. A one-line description of the task
2. Why it needs Fable specifically (not just "it's hard" — what about it requires deep reasoning, long context, cross-source synthesis, or high-stakes judgment?)
3. Estimated effort if I ran it on a cheaper model vs. Fable
4. Whether it's a one-time sprint or something I'd revisit

Rank them by impact. Flag anything that's actually NOT worth Fable so I don't waste credits after July 8.

Do not start any of the work yet. Just give me the ranked list and a one-paragraph summary of what you found in my setup that surprised you.

Key insight: The best Fable use isn't "harder chat." It's work where Claude needs to hold your whole context in its head and make judgment calls across it.

After you get the list: Pick the top 1–2 items and run Prompt 2 or 3 on them.

Prompt 2: Get life and business advice that compounds

This one turns Fable into an advisor that actually knows you. It reads your conversation history, connected apps, and stated goals, then builds a 3-month, 6-month, and 1-year plan with specific habits and milestones. Not generic productivity advice. A plan based on where you actually are.

When to use it: When you feel busy but not sure you're working on the right things. Or when you want a strategic reset before Q3.

How to run it: Paste this with Fable selected. If you have a plan doc, goals doc, or business overview, mention it. The more Claude can pull from memory and connected tools, the sharper the output.

You're my life and business advisor. I want a real assessment, not motivational fluff.

Read everything you can about my current situation:
- Previous conversations (especially anything about goals, frustrations, or big decisions)
- Any plan docs, strategy docs, or notes I've shared
- Live data from my connected apps if available (calendar, email, Slack, Drive, Notion, etc.)
- Your memory of me

Then write a one-page assessment covering:
1. Where I am right now (business, career, personal priorities — be honest, not flattering)
2. What's working that I should double down on
3. What's draining me or not moving the needle (and why)
4. The top 3 things to focus on for the next 3 months
5. What to explicitly drop or defer, and why

Then build three planning horizons:

**3-month plan:** Specific goals, weekly habits, and 2–3 milestones I can hit by October 2026.

**6-month plan:** What should be true about my business/life by January 2027 if the 3-month plan worked. Include habits that compound (not one-off tasks).

**1-year plan:** The bigger arc. Where should I be by July 2027? What systems, revenue, skills, or relationships need to be in place?

For each horizon, include:
- 3 concrete habits to start this week
- 1 thing to stop doing
- The single metric or signal that tells me I'm on track

End with: "If you only did one thing from this plan starting today, do this:" and give me the single highest-leverage move.

Be direct. I'd rather hear something uncomfortable and true than something that sounds good.

Tip: If the assessment feels too generic, follow up with: "You mentioned [specific thing from my history]. Go deeper on that and rewrite the 3-month plan around it."

Prompt 3: Plan with Opus, build with Fable

This is the big one. You have a project you've wanted to do for months (an app, a tool, a system, a site). Don't ask Fable to figure out the plan from scratch. That's expensive reasoning you don't need. Use Opus to plan. Then hand the plan to Fable to execute in one shot.

Step A — Run this on Opus (or Sonnet)

I want to plan a big project. Do not build anything yet.

Project: [describe what you want — e.g., "a client portal for my consulting business" or "a personal finance dashboard that pulls from my bank and shows monthly burn"]

Give me a full execution plan that a capable model could follow step by step without needing to re-decide anything. Include:

1. **Goal and success criteria** — What does "done" look like? How will I know it works?
2. **Phases** — Break it into ordered phases with clear deliverables for each
3. **Key decisions already made** — Lock in tech choices, scope boundaries, and design direction so the builder doesn't waffle
4. **Risks and failure modes** — What could sink this? What should the builder watch for?
5. **Open questions** — Anything you genuinely can't decide without my input (ask me now, max 3 questions)
6. **Acceptance checklist** — How to verify each phase is complete before moving on

Make the plan detailed enough that I could hand it to another person (or another model) and they'd know exactly what to build. Flag anything that's too vague to execute.

Answer any open questions Opus asks. Review the plan. Trim scope if it's too big for one session.

Step B — Run this on Fable with the plan attached

You're Fable. I'm giving you a complete execution plan created by Opus. Your job is to build this in one shot.

Rules:
- Follow the plan phase by phase. Do not re-architect or second-guess decisions already locked in unless you hit a blocker you can't resolve.
- If something in the plan is ambiguous, make a reasonable choice, note it, and keep going.
- If you hit a genuine blocker (missing API key, unclear requirement), stop and tell me exactly what you need.
- After each phase, briefly confirm what you completed against the acceptance checklist before moving on.

Here is the plan:

[Paste the full plan from Opus]

Start with Phase 1. Go.

Key insight: Opus plans cheaply. Fable executes deeply. Splitting the work this way is how you get a full project built without burning Fable credits on thinking you could've done for free.

What counts as "Fable-worthy" here: Multi-file builds, apps with real logic, systems that need to hold a lot of context, or anything where getting it 80% right on the first pass saves you a weekend of iteration.

Additional reading

Here are some related guides to check out: