How to Use Muse Spark Open in Amp
Muse Spark Open is Meta's open-weight model for debugging, software engineering, and tasks that require several reasoning steps. Its 1,048,576-token context window gives it room for extensive code and reference material.
Prompts to try with Muse Spark Open in Amp
Copy an example into Amp, or ask in your own words.
Use Muse Spark Open with high reasoning effort to trace the failing tests through this project's code. Identify the cause, make the smallest justified fix, and run the relevant tests.
Use Muse Spark Open to review the current diff for correctness. Prioritize behavior changes and missing edge cases, and cite the relevant files and lines.
Use Muse Spark Open with high reasoning effort to inspect this project's background workers for race conditions. Explain a concrete failure sequence and propose a regression test.
Use Muse Spark Open to plan this project's dependency migration. Read the dependency declarations and affected imports, then list ordered changes and checks.
Use Muse Spark Open to compare the implementation with this project's documented requirements. List discrepancies with code references and suggest focused fixes.
How to set up Muse Spark Open in Amp
Pick the way that suits you. Each takes about two minutes.
Use Spiral's MCP server
- Add the server
No header and no client ID or secret: Amp registers itself with Spiral.
amp mcp add spiral https://api.spiral.ai/mcp
- Sign in
Start amp. Amp opens Spiral's sign-in page in your browser: sign in, pick the organization and click Authorize. No API key.
- Check it
amp mcp doctor
Use the Spiral skill
- Send Amp this message
Your agent reads Spiral's skill and sets itself up.
set up https://spiral.ai/SKILL.md
- Sign in or paste a key
Approve Spiral in the browser, or give the agent a key from the console.
Use the Spiral CLI
- 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
- Create an API key
In the console, under API keys.
- Run Muse Spark Open
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": "openrouter", "model": "meta/muse-spark-open", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Amp finds Muse Spark Open on Spiral, shows the price and runs it.
Muse Spark Open providers and prices
Spiral runs Muse Spark Open through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.022688 per run at the default inputs. $1.625 per 1M input tokens, $5.525 per 1M output tokens.
Settings you can ask for
Ask Amp for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Seed | a whole number | ||
| Min p | 0 to 1 | ||
| Top k | a whole number from 0 | ||
| Top p | 0 to 1 | ||
| Logprobs | on or off | ||
| Max tokens | 1 to 943718 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Top logprobs | 0 to 20 | ||
| Presence penalty | -2 to 2 | ||
| Reasoning effort | low, medium, high, or xhigh | medium | How much the model reasons. |
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of Meta: Muse Spark Open is on its tool page: open its card above.
What can you do with Muse Spark Open in Amp?
Muse Spark Open can help carry a software task from investigation through a proposed fix and validation. Give it the relevant code, constraints, and a clear completion condition.
Trace a bug across components
Use Muse Spark Open to follow a failing request through related code, propose a cause, and identify a test that would confirm it.
Plan a software migration
Ask Muse Spark Open to turn requirements and compatibility constraints into ordered changes, validation checks, and rollback points.
Review a patch for correctness
Have Muse Spark Open examine a patch for behavior changes, missing cases, and assumptions that need testing.
Compare technical proposals
Give Muse Spark Open competing designs and explicit constraints, then request tradeoffs and unresolved questions.
Triage incoming reports
Use Muse Spark Open with low reasoning effort to classify reports, extract reproduction details, and flag missing evidence.
Investigate a failed workflow
Ask Muse Spark Open to reason through successive tool results, locate the failed step, and propose a recovery plan.
More on Muse Spark Open and Amp
Overview
Muse Spark Open in other agents
Questions about Muse Spark Open in Amp
Can Amp use Muse Spark Open?
Yes. Add Spiral to Amp as an MCP server, and Amp can find and run Muse Spark Open like any built-in tool.
How much does Muse Spark Open cost in Amp?
Up to $0.022688 per run at the default inputs. $1.625 per 1M input tokens, $5.525 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Amp plan is separate.
Do I need an OpenRouter account to use Muse Spark Open in Amp?
No. Spiral runs Muse Spark Open on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up Muse Spark Open in Amp?
About two minutes: add the server, then sign in, then check it.
Does Amp wait while Muse Spark Open runs?
Muse Spark Open runs in the background. Amp starts the run, checks on it and gives you the text, json when it's ready.
What is Muse Spark Open best suited for?
Meta designed Muse Spark Open for coding and agentic work. Useful tasks include debugging across several components, planning software changes, and working through problems that require multiple steps.
How much context can Muse Spark Open handle?
Meta specifies a 1,048,576-token context window. That is the context limit, not a promise of an equally long answer.
Can I turn off reasoning?
No. Muse Spark Open always reasons before answering; use a lower reasoning effort for simpler requests.
Which reasoning effort should I choose?
Use low for extraction or classification, medium for general agent tasks, and high for difficult debugging or multi-step problems. Choose xhigh when a deeper attempt matters more than response speed.
Does Muse Spark Open support languages besides English?
Yes. Meta says it was trained on data from more than 100 languages; test representative examples in your target language before relying on the results.
Start using Muse Spark Open in Amp
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into Amp:
set up https://spiral.ai/SKILL.md


