[Integrations](https://spiral.ai/integrations) › [dots](https://spiral.ai/integrations/dots) › [Claude Haiku Latest](https://spiral.ai/integrations/dots/claude-haiku-latest)

# How to Use Claude Haiku Latest in dots

Use Claude Haiku Latest for quick classification, extraction, summaries, and focused coding tasks. Anthropic currently lists Claude Haiku 5.5, which expands context and output limits compared with Claude Haiku 4.5.

|  |  |
| --- | --- |
| Model | Claude Haiku Latest |
| Available through | Anthropic |
| Category | LLM |
| Price | Up to $0.002672 per run at the default inputs. $0.13 per 1M input tokens, $0.013 per 1M cached input tokens, $0.1625 per 1M cache write tokens, $0.26 per 1M 1-hour cache write tokens, $0.65 per 1M output tokens. |
| Returns | text, json |
| Runs | In the background |
| Setup | About 2 minutes, with a Spiral API key |

## Prompts to try with Claude Haiku Latest in dots

- Use Claude Haiku Latest on each scheduled run to classify newly supplied support tickets as billing, access, bug, or feature request. Include ticket identifiers and supporting excerpts; mark ambiguous tickets for review.
- Use Claude Haiku Latest to summarize the activity updates supplied with today's run in under 250 words. Group decisions, blockers, and next actions, and retain the original update identifiers.
- Use Claude Haiku Latest to extract order numbers, reported problems, and requested actions from the customer messages supplied with this run. Leave missing details empty and include the message identifier for every record.
- Use Claude Haiku Latest to draft replies for the unanswered questions supplied with this run using the supplied policy text. Keep each draft under 100 words and flag policy gaps for review.
- Use Claude Haiku Latest to create handoff notes for the completed support conversations supplied with this run. For each conversation, list the goal, confirmed facts, unresolved questions, and next action.

## How to set up Claude Haiku Latest in dots

Pick the way that suits you. Each takes about two minutes.

### Use the Spiral skill

1. **Send dots this message**: Your agent reads Spiral's skill and sets itself up.

```
set up https://spiral.ai/SKILL.md
```

2. **Sign in or paste a key**: Approve Spiral in the browser, or give the agent a key from the console.

### Use Spiral's MCP server

1. **Create an API key**: In the console, open API keys and create a key for your dot. Copy it.

2. **Send your dot this message**: Your dot reads Spiral's skill and asks for your API key.

```text
set up https://spiral.ai/SKILL.md
```

3. **Give your dot the key**: When your dot asks for the key, add it where your dot keeps secrets, never in the chat itself.

### Use the Spiral CLI

1. **Install and sign in**: Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run Claude Haiku Latest**: One request; the result comes back by polling or a webhook.

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "anthropic",
    "model": "~anthropic/claude-haiku-latest",
    "input": {
      "messages": [
        {
          "role": "system"
        }
      ]
    }
  }'
```

Then paste one of the prompts above: dots finds Claude Haiku Latest on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/dots

## Claude Haiku Latest tools and prices

- [Claude Haiku Latest](https://spiral.ai/tools/anthropic/~anthropic/claude-haiku-latest) (Anthropic): Up to $0.002672 per run at the default inputs. $0.13 per 1M input tokens, $0.013 per 1M cached input tokens, $0.1625 per 1M cache write tokens, $0.26 per 1M 1-hour cache write tokens, $0.65 per 1M output tokens.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| stop |  | — |  |
| tools |  | — |  |
| reasoning |  | — | Reasoning effort; reasoning tokens count toward max_tokens. |
| max_tokens | 1 to 128000 | — | The most output tokens to generate. On a reasoning model the reasoning tokens count toward it: a cap too low for the effort can end with no answer (finish_reason length). |
| tool_choice |  | — |  |
| response_format |  | — |  |

## What can you do with Claude Haiku Latest in dots?

Claude Haiku can sort recurring batches of messages and produce short reports from supplied updates. Define the categories, evidence requirements, and cases that need human review.

### Support ticket routing

Classify incoming tickets as billing, access, bug, or feature request. Ask for a short supporting excerpt and send ambiguous cases to review.

### Extracting details from messages

Turn customer messages into records containing order numbers, reported problems, and requested actions. Require null for details the message does not supply.

### Daily activity summaries

Condense supplied updates into decisions, blockers, and next actions. Keep the original item identifiers beside each point for checking.

### Customer support drafts

Draft short replies using supplied policy text and conversation history. Ask for a review flag when the policy does not answer the customer's question.

### Database query drafts

Write read-only SQL from a supplied schema and a plain-English question. Require an explanation of joins and assumptions before reviewing the query.

### Focused code fixes

Assign a reproducible bug or a small refactor with clear acceptance criteria. Require a check that exercises the changed behavior before accepting the result.

### Multilingual support content

Translate help text or draft replies in a specified language. Supply a glossary and ask the model to preserve product names and technical terms.

### Conversation handoff notes

Summarize a supplied conversation into the customer's goal, established facts, unresolved questions, and next action. Preserve exact identifiers needed by the next person.

## Overview

- [Anthropic Claude](https://spiral.ai/providers/anthropic)
- [All dots guides](https://spiral.ai/integrations/dots)

## Claude Haiku Latest in other agents

- [Amp](https://spiral.ai/integrations/amp/claude-haiku-latest)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/claude-haiku-latest)
- [Claude](https://spiral.ai/integrations/claude-desktop/claude-haiku-latest)
- [Claude Code](https://spiral.ai/integrations/claude-code/claude-haiku-latest)
- [Cline](https://spiral.ai/integrations/cline/claude-haiku-latest)
- [Codex](https://spiral.ai/integrations/codex/claude-haiku-latest)
- [Cursor](https://spiral.ai/integrations/cursor/claude-haiku-latest)
- [Devin](https://spiral.ai/integrations/devin/claude-haiku-latest)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/claude-haiku-latest)
- [Gemini](https://spiral.ai/integrations/gemini/claude-haiku-latest)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/claude-haiku-latest)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/claude-haiku-latest)
- [goose](https://spiral.ai/integrations/goose/claude-haiku-latest)
- [Grok](https://spiral.ai/integrations/grok/claude-haiku-latest)
- [Grok Build](https://spiral.ai/integrations/grok-build/claude-haiku-latest)
- [Hermes](https://spiral.ai/integrations/hermes/claude-haiku-latest)
- [Instinct](https://spiral.ai/integrations/instinct/claude-haiku-latest)
- [Manus](https://spiral.ai/integrations/manus/claude-haiku-latest)
- [Muse](https://spiral.ai/integrations/muse/claude-haiku-latest)
- [Muse Code](https://spiral.ai/integrations/muse-code/claude-haiku-latest)
- [n8n](https://spiral.ai/integrations/n8n/claude-haiku-latest)
- [Notion](https://spiral.ai/integrations/notion/claude-haiku-latest)
- [OpenClaw](https://spiral.ai/integrations/openclaw/claude-haiku-latest)
- [opencode](https://spiral.ai/integrations/opencode/claude-haiku-latest)
- [Perplexity](https://spiral.ai/integrations/perplexity/claude-haiku-latest)
- [Raycast](https://spiral.ai/integrations/raycast/claude-haiku-latest)
- [VS Code](https://spiral.ai/integrations/vscode/claude-haiku-latest)
- [Windsurf](https://spiral.ai/integrations/windsurf/claude-haiku-latest)
- [Zapier](https://spiral.ai/integrations/zapier/claude-haiku-latest)

## More dots guides

- [Claude Haiku 5.5](https://spiral.ai/integrations/dots/claude-haiku-5-5)
- [Claude Opus 4.5](https://spiral.ai/integrations/dots/claude-opus-4-5)
- [Claude Haiku 4.5](https://spiral.ai/integrations/dots/claude-haiku-4-5)
- [Claude Opus 4.6](https://spiral.ai/integrations/dots/claude-opus-4-6)
- [Claude Fable Latest](https://spiral.ai/integrations/dots/claude-fable-latest)
- [Claude Opus 4.7](https://spiral.ai/integrations/dots/claude-opus-4-7)

## Questions about Claude Haiku Latest in dots

**Can dots use Claude Haiku Latest?**

Yes. Add Spiral to dots as an MCP server, and dots can find and run Claude Haiku Latest like any built-in tool.

**How much does Claude Haiku Latest cost in dots?**

Up to $0.002672 per run at the default inputs. $0.13 per 1M input tokens, $0.013 per 1M cached input tokens, $0.1625 per 1M cache write tokens, $0.26 per 1M 1-hour cache write tokens, $0.65 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your dots plan is separate.

**Do I need an Anthropic account to use Claude Haiku Latest in dots?**

You need no Anthropic account. dots uses one Spiral API key for every tool, and Spiral runs Claude Haiku Latest on its own Anthropic account.

**How long does it take to set up Claude Haiku Latest in dots?**

About two minutes: create an API key, then send your dot this message, then give your dot the key.

**Does dots wait while Claude Haiku Latest runs?**

Claude Haiku Latest runs in the background. dots starts the run, checks on it and gives you the text, json when it's ready.

**What is Claude Haiku good at?**

Claude Haiku is useful for repeated, clearly defined tasks such as classifying requests, extracting details, drafting summaries, and answering support questions. Focused coding tasks are another practical use.

**How much text can Claude Haiku handle?**

Anthropic currently lists Claude Haiku 5.5 with a 1 million token context window and up to 128,000 output tokens. Claude Haiku 4.5 has a 200,000 token context window and a 64,000 token output limit, so keep limits tied to the named release.

**Which reasoning effort should I use?**

For Claude Haiku 5.5, start with medium effort for most tasks, including coding. Try high effort for longer assignments, knowledge work, or tasks where following every instruction matters.

**What should I watch for when using Claude Haiku for code?**

Claude Haiku 5.5 sometimes reports a change complete without running a check at low or medium effort. Explicitly ask it to run a test, build, or type check that exercises the change and report the result.

**Can Claude Haiku work in languages besides English?**

Yes. Anthropic lists multilingual capabilities for its current models, including Claude Haiku 5.5; specify the target language, audience, and terms that must stay unchanged.

