Claude in Whittl

Anthropic's Claude is Whittl's highest-quality backend overall — consistently strong on complex Python code, excellent at following multi-step instructions, native vision across all tiers, and the most reliable tool-use implementation.

The tradeoff: most expensive per token of any backend. For serious work where iteration count matters more than per-call cost, it's the default pick.

Getting a key

1. Go to console.anthropic.com.
2. Sign up / log in.
3. API Keys in the left sidebar → Create Key.
4. Name it "Whittl", copy the key (starts with sk-ant-).

Add it in Whittl from the key button next to the backend dropdown in the chat panel, in the Claude API Key field. The first-run setup wizard asks for it too.

Picking a model

Four tiers since v2.5.0, each with a clear role. The dropdown next to the backend selector lists them as Haiku 4.5 (cheap), Sonnet 5, Opus 5 (smart) and Fable 5 (max). Each tier has a fallback chain, so if Anthropic retires a model id Whittl steps down to the previous generation rather than failing.

Fable 5 (max)

Anthropic's most capable generally available model, with 1M context and 128K output. It always thinks; there is no way to turn reasoning off, and Whittl sends the request accordingly. Roughly twice the price of Opus. Use it for the hardest architectural problems and for Agent Mode tasks where a single wrong turn costs more than the model does. For everything else it is overkill.

Opus 5 (smart)

The quality ceiling for routine premium work: 1M context, 128K output, full effort ladder for thinking depth. Use Opus for complex refactors, long-context tasks, and difficult multi-round Agent Mode work.

When to NOT use Opus: simple edits and quick iterations, where Sonnet handles it at a fraction of the cost.

Sonnet 5 (default pick)

The "just use Claude" model: 1M context, 128K output, most of Opus's quality at a much lower price. For most Whittl work, Sonnet is the right choice.

• Generating new apps from natural language
• Modifying existing projects
• Screenshot to App
• Most Agent Mode tasks

Haiku 4.5 (cheap)

Fast, cheap, surprisingly capable, and it supports extended thinking. Use Haiku for small surgical edits and quick iterations on an established project; expect it to fall behind Sonnet on complex architecture and visual design. Field-measured across a month of real sessions: about $0.32 per session with caching.

Pricing changes often enough that this page does not quote per-token rates. Check anthropic.com/pricing; the cost meter in Whittl's status bar shows what each generation actually cost, including cached tokens.

Whittl adapts each request to the model: parameters the newer models reject are never sent, and a rejected parameter no longer causes a silent fallback to a cheaper tier, which was a real failure mode before v2.5.0.

Switching models

The Claude model dropdown next to the backend selector chooses which Claude variant to use. Settings persist per project.

You can also switch mid-session — conversation history carries forward. Common pattern: start a project on Sonnet, iterate most changes with Haiku, escalate one hard problem back to Sonnet or Opus.

Prompt caching

Claude supports prompt caching: the first request pays full price for the system prompt + tool definitions + code context; subsequent requests in the same session pay ~10% of that. The savings compound across multi-round sessions.

In practice this means:

• A fresh project on Sonnet: first request ~$0.15, later requests ~$0.03
• Long Agent Mode session (10+ rounds): ~87% cheaper than without caching

Whittl enables caching automatically. You don't need to configure anything — just use Claude and reap the savings.

Vision

Every Claude tier supports vision natively via Anthropic's image content blocks. Drop a screenshot into chat and the AI sees it.

Compared to OpenRouter → Claude via OpenAI-format multimodal:

• Direct Claude API: native Anthropic image block format. Slightly richer image reasoning.
• OpenRouter → Claude: OpenAI-format image_url content parts. Close but not identical.

If you care about the highest image-reasoning fidelity on a specific project, use the Claude API backend directly rather than Claude via OpenRouter.

Tool use

Claude's tool-use API is the most reliable in practice. Whittl's edit_code, edit_function, multi_edit, syntax_check, create_file, glob, run_code, and (in Agent Mode) bash, web_search and web_fetch tools all fire cleanly on Claude. See The AI's Tools for the full list. The native text_editor_20250728 tool is used on Claude specifically for multi-round surgical editing.

Other backends work fine — Claude just produces fewer tool-call format errors or hallucinated tool arguments on edge cases.

Rate limits

Anthropic's rate limits scale with your account tier:

• Free tier: very restrictive, usually maxes out quickly
• Usage tier 1 ($5 spent): 50 requests/min, 50K tokens/min
• Usage tier 2, 3, 4: higher limits as you spend more

If you hit rate limits, Whittl surfaces the error and you can either wait or switch to OpenRouter → Claude (which pools requests across OpenRouter's user base).

Claude pricing and cost notes

Use Haiku by default

Switch to Haiku for most work, escalate to Sonnet when quality matters. Most sessions can finish on Haiku alone.

Let prompt caching do its job

Keep sessions long. Starting a fresh project for every task defeats caching.

Multi-file projects benefit disproportionately

Smart routing + caching means your 30-file project costs roughly the same per edit as a single-file one, because the cached context covers the common ~80% and each new prompt only pays for the delta.

When NOT to use Claude

• Budget-only work. Claude Haiku is cheap but DeepSeek or OpenRouter free-tier models are cheaper.
• Offline / privacy-critical work. Use Ollama.
• Exploratory model comparison. Use OpenRouter to swap between many models.

Troubleshooting

Invalid API key when the key looks right

Common causes: you copied whitespace into the key, or your Anthropic account is new and hasn't had its first credit purchase yet. Re-copy the key with a trimmed edit, and check your account has a valid payment method.

Model consistently slow

Anthropic's capacity fluctuates by time of day. US peak hours (9am-5pm Pacific) are slower than off-peak. Not a Whittl issue — happens on every tool using the Claude API.

Seeing 529 Overloaded errors

Anthropic's load shedding. Whittl retries automatically with backoff; if it keeps failing, wait 5-10 minutes or switch to a different backend for that session.

What's next

• Agent Mode — where Claude's tool-use and long-context strengths pay off most
• OpenRouter — alternative path to Claude if you prefer consolidated billing
• Choosing a Backend — comparison with other backends