How to Use DeepSeek V3.1 in Amp
Use DeepSeek V3.1 for code fixes, research with search tools, and complex reasoning tasks. It combines thinking and non-thinking responses, with standard tool calling documented for non-thinking mode.
Prompts to try with DeepSeek V3.1 in Amp
Copy an example into Amp, or ask in your own words.
Use DeepSeek V3.1 in non-thinking mode to investigate the failing tests in this project. Reproduce one failure, inspect the relevant implementation, make the smallest justified fix, and run the affected tests.
Use DeepSeek V3.1 to review the current diff. Report only findings with a concrete failure scenario, cite file locations, and propose a test for each finding.
Use DeepSeek V3.1 with thinking enabled to compare the retry logic in this project with bounded exponential backoff. Explain the tradeoffs and recommend a design before editing.
Use DeepSeek V3.1 in non-thinking mode to add regression tests for the parser's malformed-input handling. Read the existing tests first and verify observable behavior rather than internal implementation details.
Use DeepSeek V3.1 to compare this project's README with its public interfaces. Identify claims the code does not support, propose precise documentation changes, and cite the relevant implementation.
How to set up DeepSeek V3.1 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 DeepSeek V3.1
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": "deepseek/deepseek-chat-v3.1", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Amp finds DeepSeek V3.1 on Spiral, shows the price and runs it.
DeepSeek V3.1 providers and prices
Spiral runs DeepSeek V3.1 through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.00507 per run at the default inputs. $0.325 per 1M input tokens, $1.235 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 147456 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Top logprobs | 0 to 20 | ||
| Presence penalty | -2 to 2 | ||
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of DeepSeek: DeepSeek V3.1 is on its tool page: open its card above.
What can you do with DeepSeek V3.1 in Amp?
Turn project code and test failures into proposed patches with DeepSeek V3.1. Give the agent relevant files, expected behavior, and a clear verification target.
Debug a failing test
Give DeepSeek V3.1 the failure, relevant code, and expected behavior. Ask for a minimal patch and a regression test that reproduces the bug.
Review a proposed code change
Ask DeepSeek V3.1 to inspect a diff for behavior changes, missing error handling, and test gaps. Request a concrete failure scenario for each finding.
Compare technical designs
Use DeepSeek V3.1 to compare implementation approaches against stated constraints. Request a recommendation, assumptions, and conditions that would change the decision.
Research a technical question
Give DeepSeek V3.1 access to a search tool and ask it to investigate a question across several sources. Require dated evidence and links for the final conclusions.
Compare long specifications
Provide specification text within the documented 128K context window. Ask DeepSeek V3.1 to identify conflicting requirements and quote the passages behind each conflict.
Triage operational errors
Supply error logs and incident notes as text. Ask DeepSeek V3.1 to group recurring failures, rank possible causes, and propose checks that distinguish them.
Turn support questions into technical investigations
Ask DeepSeek V3.1 to compare a reported problem with supplied documentation, identify missing evidence, and draft focused diagnostic questions.
More on DeepSeek V3.1 and Amp
Overview
DeepSeek V3.1 in other agents
Questions about DeepSeek V3.1 in Amp
Can Amp use DeepSeek V3.1?
Yes. Add Spiral to Amp as an MCP server, and Amp can find and run DeepSeek V3.1 like any built-in tool.
How much does DeepSeek V3.1 cost in Amp?
Up to $0.00507 per run at the default inputs. $0.325 per 1M input tokens, $1.235 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 DeepSeek V3.1 in Amp?
No. Spiral runs DeepSeek V3.1 on its own OpenRouter account. You sign in to Spiral once and never paste a key.
How long does it take to set up DeepSeek V3.1 in Amp?
About two minutes: add the server, then sign in, then check it.
Does Amp wait while DeepSeek V3.1 runs?
DeepSeek V3.1 runs in the background. Amp starts the run, checks on it and gives you the text, json when it's ready.
What is DeepSeek V3.1 useful for?
Code generation, tool use, and reasoning tasks are its main targets. Try it for debugging, research with search tools, or comparing technical approaches against explicit constraints.
When should I use thinking mode?
DeepSeek V3.1 supports thinking and non-thinking responses. Try thinking mode for problems with several constraints or competing explanations, and non-thinking mode for direct drafting or extraction.
How much context does DeepSeek V3.1 support?
DeepSeek documents a 128K context window in both modes. For long material, specify the sections to compare and the evidence you want quoted.
Can DeepSeek V3.1 research current information?
Yes, when supplied with a search tool. DeepSeek documents a process that searches across multiple turns; ask for dated sources and links supporting the final claims.
Can DeepSeek V3.1 call tools while thinking?
DeepSeek documents standard tool calling in non-thinking mode. It also documents a separate search-agent format for thinking mode, so ordinary tool calling and search during reasoning should be treated separately.
Start using DeepSeek V3.1 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


