Spiral gives your agent every tool with one account. Browse the tools →

OpenClaw · LLMReleased 2026-05-28

How to Use Claude Opus 4.8 in OpenClaw

Claude Opus 4.8 suits complex coding, document analysis, and tasks that require several steps. Compared with Claude Opus 4.7, Anthropic reports it is about four times less likely to leave flaws in its own code unflagged.

Prompts to try with Claude Opus 4.8 in OpenClaw

Copy an example into OpenClaw, or ask in your own words.

Every weekday, use Claude Opus 4.8 to analyze the incident summaries supplied for that day. Return recurring causes, unresolved issues, and supporting incident references in no more than 500 words.

Claude Opus 4.8OpenClaw
Try asking
  • Each Friday, use Claude Opus 4.8 with high reasoning effort to compare the supplied weekly project updates. Report conflicting dates, dependencies without owners, and decisions requiring review.

  • Every morning, use Claude Opus 4.8 to extract commitments from the newly supplied correspondence. Include the owner, deadline, and source passage for every commitment, and mark missing details as unknown.

  • Each week, use Claude Opus 4.8 to synthesize the supplied research excerpts. Organize the brief into new findings, disagreements, and unanswered questions, with a source reference for every finding.

  • At the end of each day, use Claude Opus 4.8 to analyze the supplied support conversation text. Group recurring product problems and give a representative conversation reference for each.

How to set up Claude Opus 4.8 in OpenClaw

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

Use the Spiral skill

  1. Send OpenClaw 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. Add the server

    On the machine that runs OpenClaw's Gateway:

    openclaw mcp add spiral \
      --url https://api.spiral.ai/mcp \
      --transport streamable-http \
      --auth oauth
  2. Sign in

    In Settings › MCP, select Sign in on the Spiral server, sign in to Spiral and click Authorize. No API key.

    openclaw mcp login spiral
  3. Reload and check

    A running Gateway may keep its old tool list: restart it, or Save & Publish in the Control UI.

    openclaw mcp reload
    openclaw mcp doctor spiral --probe

Use the Spiral CLI

  1. Install and sign in

    Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

    npx -y spiral-cli@latest init

Call the Spiral API

  1. Create an API key

    In the console, under API keys.

  2. Run Claude Opus 4.8

    One request; the result comes back by polling or a webhook.

    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-opus-4.8",
        "input": {
          "messages": [
            {
              "role": "system"
            }
          ]
        }
      }'

Then paste one of the prompts above: OpenClaw finds Claude Opus 4.8 on Spiral, shows the price and runs it.

Claude Opus 4.8 providers and prices

Spiral runs Claude Opus 4.8 through Anthropic. Open a card for today's price, with every fee included.

AnthropicLLM
Claude Opus 4.8
anthropic/claude-opus-4.8

Settings you can ask for

Ask OpenClaw for any of these in your prompt.

SettingAllowed valuesDefaultDescription
Max tokens1 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).

Reasoning effortnone, minimal, low, medium, high, xhigh, or maxnone

How much the model reasons.

Parallel tool callson or off

Every API field of Claude Opus 4.8 is on its tool page: open its card above.

What can you do with Claude Opus 4.8 in OpenClaw?

Claude Opus 4.8 can turn recurring batches of text into consistent reports and records. Define the reporting period, selection criteria, and required evidence so each run produces comparable results.

Investigate a difficult bug

Trace a failure across related functions, identify the cause, and propose a fix with a regression check.

Review a code change

Look for incorrect behavior, missing error handling, and test failures. Request evidence, severity, and confidence for each finding.

Plan a code migration

Compare existing behavior with the target requirements, map dependencies, and produce an ordered implementation and validation plan.

Compare competing proposals

Evaluate supplied proposals against the same criteria and produce a decision table that exposes missing information.

Extract records from text

Turn meeting notes or business correspondence into consistently formatted records with explicit rules for missing values.

Build a research brief

Combine supplied source text into a focused brief that separates agreement, conflicting evidence, and unanswered questions.

Draft a technical decision memo

Turn requirements, constraints, and proposed approaches into a memo covering tradeoffs and the evidence needed to choose.

Summarize recurring operational issues

Analyze a supplied batch of incident summaries or support conversations and organize recurring problems for review.

Questions about Claude Opus 4.8 in OpenClaw

Can OpenClaw use Claude Opus 4.8?

Yes. Add Spiral to OpenClaw as an MCP server, and OpenClaw can find and run Claude Opus 4.8 like any built-in tool.

How much does Claude Opus 4.8 cost in OpenClaw?

Up to $0.133575 per run at the default inputs. $6.50 per 1M input tokens, $0.65 per 1M cached input tokens, $8.125 per 1M cache write tokens, $13.00 per 1M 1-hour cache write tokens, $32.50 per 1M output tokens; for a prompt above 200,000 tokens, $13.00 per 1M input tokens, $1.30 per 1M cached input tokens, $16.25 per 1M cache write tokens, $26.00 per 1M 1-hour cache write tokens, $48.75 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your OpenClaw plan is separate.

Do I need an Anthropic account to use Claude Opus 4.8 in OpenClaw?

No. Spiral runs Claude Opus 4.8 on its own Anthropic account. You sign in to Spiral once and never paste a key.

How long does it take to set up Claude Opus 4.8 in OpenClaw?

About two minutes: add the server, then sign in, then reload and check.

Does OpenClaw wait while Claude Opus 4.8 runs?

Claude Opus 4.8 runs in the background. OpenClaw starts the run, checks on it and gives you the text, json when it's ready.

What is Claude Opus 4.8 good at?

Use it for complex coding, bug investigation, document analysis, and work that takes several steps. It also suits extracting consistent records from text when you define the expected output precisely.

How much text can Claude Opus 4.8 handle?

Its context window holds up to 1 million tokens, including conversation history and generated output, with a standard output limit of 128,000 tokens. Keep the supplied material relevant because accuracy and recall can decline as context grows.

Which reasoning effort should I use?

Anthropic recommends extra high effort for coding and tasks involving multiple agent actions, and high effort for other demanding analysis. Try lower effort on narrowly defined tasks and check whether the results still meet your requirements.

How should I prompt Claude Opus 4.8?

State the task, scope, constraints, and exact output you want. Say whether each instruction applies to every record or section because Claude Opus 4.8 follows wording literally.

Why might Claude Opus 4.8 omit bugs during a code review?

Instructions such as “only report important issues” can make it filter out bugs it considers below that threshold. Define what counts as a reportable bug, or request findings with severity and confidence so you can assess them afterward.

Get started

Start using Claude Opus 4.8 in OpenClaw

One Spiral account for every tool your agent needs. No subscription; you pay per run.

Paste this into OpenClaw:

Set up OpenClaw

set up https://spiral.ai/SKILL.md
Are you an AI agent? See setup options