How to Use Inkling in Amp
Inkling suits coding, tool-driven tasks, and answers grounded in supplied documents. Its adjustable thinking effort lets you balance reasoning depth against token use.
Prompts to try with Inkling in Amp
Copy an example into Amp, or ask in your own words.
Use Inkling to inspect this project's failing tests, identify the cause, and make the smallest justified fix. Add a regression check and report what you ran.
Use Inkling to build a searchable release-notes page using this project's existing components. Include keyboard navigation, an empty state, and tests for the search behavior.
Use Inkling with more thinking effort to trace how authentication state moves through this repository. Cite file locations, identify inconsistent behavior, and propose a focused repair.
Use Inkling to compare the API implementation with the specification in this project. List discrepancies with file references, then fix the discrepancies covered by existing requirements.
Use Inkling to review this diff for behavior changes and missing error handling. Explain each finding with a concrete failure scenario and suggest a test that would catch it.
How to set up Inkling 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 Inkling
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": "thinkingmachines/inkling", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Amp finds Inkling on Spiral, shows the price and runs it.
Inkling providers and prices
Spiral runs Inkling through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.021612 per run at the default inputs. $1.30 per 1M input tokens, $5.265 per 1M output tokens.
Settings you can ask for
Ask Amp for any of these in your prompt.
| Setting | Allowed values | Default | Description |
|---|---|---|---|
| Min p | 0 to 1 | ||
| Top k | a whole number from 0 | ||
| Top p | 0 to 1 | ||
| Max tokens | 1 to 471859 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Presence penalty | -2 to 2 | ||
| Reasoning effort | none, minimal, low, medium, high, or max | high | How much the model reasons. |
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of Thinking Machines: Inkling is on its tool page: open its card above.
What can you do with Inkling in Amp?
Turn a bug report or feature specification into code with Inkling's coding and reasoning capabilities. Give the agent project context and measurable acceptance criteria so it can check the result.
Fix a reproducible software bug
Give Inkling the failing behavior, relevant code, and expected result. Ask for a focused patch and a regression check that reproduces the original failure.
Build a small web application
Ask Inkling to implement a concrete workflow with explicit requirements, such as a form with validation and a results view. Define what must work before requesting visual refinements.
Answer questions from a knowledge base
Supply relevant passages and ask Inkling to cite passage identifiers in its answer. Require an explicit statement when the supplied material does not answer the question.
Compare conflicting documents
Have Inkling compare policy drafts or product specifications, identify contradictions, and link each finding to the supporting passage. Keep proposed resolutions separate from documented requirements.
Draft multilingual support replies
Give Inkling the customer's message, approved support guidance, and desired response language. Ask it to preserve product names and flag missing policy details.
Investigate a technical incident
Provide a timeline and relevant log excerpts. Ask Inkling to rank possible causes, cite evidence, and suggest a check that could disprove each explanation.
Prepare a recurring change digest
Use Inkling to compare newly retrieved release notes with the previous report. Request a short digest of changed behavior, supporting links, and unresolved questions.
More on Inkling and Amp
Overview
Inkling in other agents
Questions about Inkling in Amp
Can Amp use Inkling?
Yes. Add Spiral to Amp as an MCP server, and Amp can find and run Inkling like any built-in tool.
How much does Inkling cost in Amp?
Up to $0.021612 per run at the default inputs. $1.30 per 1M input tokens, $5.265 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 Inkling in Amp?
No. Spiral runs Inkling on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up Inkling in Amp?
About two minutes: add the server, then sign in, then check it.
Does Amp wait while Inkling runs?
Inkling runs in the background. Amp starts the run, checks on it and gives you the text, json when it's ready.
What is Inkling good for?
Inkling is designed for coding, reasoning, tool use, and answering questions from retrieved information. Practical tasks include fixing software bugs, comparing documents, and drafting answers grounded in a knowledge base.
How much context can Inkling handle?
Thinking Machines documents up to one million tokens of model context. Treat that as the model's documented capacity, rather than a guarantee that every request can use the full window.
When should I increase Inkling's thinking effort?
Try more thinking effort for difficult debugging, competing explanations, or problems with several constraints. Compare the results on representative tasks: the control balances performance against token use.
What are Inkling's weaknesses?
Inkling can produce plausible but unsupported statements, and its maker does not claim it is the strongest overall model. For factual tasks, supply evidence and ask it to distinguish supported conclusions from assumptions.
Can Inkling work in languages other than English?
Yes. Thinking Machines describes general multilingual capabilities; test your target language and terminology with examples before using it for repeated work.
Start using Inkling 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


