How to Use DeepSeek V3 in Instinct
DeepSeek V3 is a practical choice for coding, math, and drafting from supplied text. Its published 128K context supports working through long passages and code excerpts in one conversation.
Prompts to try with DeepSeek V3 in Instinct
Copy an example into Instinct, or ask in your own words.
Each morning, use DeepSeek V3 to summarize the supplied support conversations in under 400 words. Group recurring problems, quote supporting messages, and separate customer reports from confirmed causes.
For each incoming feedback batch, use DeepSeek V3 to return JSON in this structure: {"items":[{"topic":"login","summary":"Cannot sign in","evidence":"Customer reports a login error"}]}. Include only supplied feedback and limit each batch to 20 items.
Each week, use DeepSeek V3 to compare the supplied current and previous policy text. List changed obligations with a supporting excerpt from each version, and flag ambiguous changes.
For each supplied product announcement, use DeepSeek V3 to draft an English and a Simplified Chinese summary, each under 150 words. Preserve names and numbers and follow the supplied glossary.
Every Friday, use DeepSeek V3 to turn the supplied release notes into a customer update under 300 words. Separate fixes from new capabilities and include only claims supported by the notes.
How to set up DeepSeek V3 in Instinct
Pick the way that suits you. Each takes about two minutes.
Use the Spiral skill
- Send Instinct 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 Spiral's MCP server
- Create an API key
In the console, open API keys and create a key for Instinct. Copy it.
- Send Instinct this message
Instinct reads Spiral's skill and asks for your API key.
set up https://spiral.ai/SKILL.md
- Give Instinct the key
When Instinct asks for the key, add it where Instinct keeps secrets, never in the chat itself.
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
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", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Instinct finds DeepSeek V3 on Spiral, shows the price and runs it.
DeepSeek V3 providers and prices
Spiral runs DeepSeek V3 through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.00549 per run at the default inputs. $0.33462 per 1M input tokens, $1.33731 per 1M output tokens.
Settings you can ask for
Ask Instinct 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 | ||
| Max tokens | 1 to 16384 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Presence penalty | -2 to 2 | ||
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of DeepSeek: DeepSeek V3 is on its tool page: open its card above.
What can you do with DeepSeek V3 in Instinct?
Turn recurring batches of text into reviewable summaries, translations, and structured records with DeepSeek V3. Define the evidence requirements and output format once, then apply them to each incoming batch.
Explain unfamiliar code
Ask DeepSeek V3 to trace a function's inputs, branches, and outputs, then identify assumptions that deserve tests.
Draft a bug fix
Give DeepSeek V3 the failing behavior, relevant code, and expected result. Request a small patch and a regression test for review.
Check a quantitative argument
Use DeepSeek V3 to work through a calculation with units, explicit assumptions, and an independent check of the result.
Compare long reference passages
Supply labeled passages to DeepSeek V3 and request a comparison with supporting excerpts and unresolved questions.
Extract structured records
Have DeepSeek V3 turn supplied text into JSON using an example object. Require null for missing information and validate each result.
Prepare English and Chinese drafts
Ask DeepSeek V3 to translate supplied text between English and Chinese while preserving a glossary, names, and numerical details.
Practice a difficult conversation
Use DeepSeek V3 to role-play a customer interview or support exchange, then review which questions uncovered useful information.
Draft short fiction
Give DeepSeek V3 a character, conflict, and word limit. Request a complete scene, then revise it against specific editorial constraints.
More on DeepSeek V3 and Instinct
Overview
DeepSeek V3 in other agents
Questions about DeepSeek V3 in Instinct
Can Instinct use DeepSeek V3?
Yes. Add Spiral to Instinct as an MCP server, and Instinct can find and run DeepSeek V3 like any built-in tool.
How much does DeepSeek V3 cost in Instinct?
Up to $0.00549 per run at the default inputs. $0.33462 per 1M input tokens, $1.33731 per 1M output tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Instinct plan is separate.
Do I need an OpenRouter account to use DeepSeek V3 in Instinct?
You need no OpenRouter account. Instinct uses one Spiral API key for every tool, and Spiral runs DeepSeek V3 on its own OpenRouter account.
How long does it take to set up DeepSeek V3 in Instinct?
About two minutes: create an API key, then send Instinct this message, then give Instinct the key.
Does Instinct wait while DeepSeek V3 runs?
DeepSeek V3 runs in the background. Instinct starts the run, checks on it and gives you the text, json when it's ready.
What is DeepSeek V3 good at?
Coding and math stand out in DeepSeek's reported evaluations. Practical uses include explaining code, proposing fixes, and working through quantitative problems with explicit assumptions.
How much context can DeepSeek V3 handle?
DeepSeek's original model documentation reports a 128K-token context and retrieval tests through that length. Use that as the published model reference, and request a focused answer when supplying long material.
Which languages does DeepSeek V3 support?
DeepSeek V3 was trained on multilingual text, with English and Chinese making up the majority. For translation work, specify the target language, audience, and terms that must stay unchanged.
Can I trust DeepSeek V3's factual answers?
Verify factual answers against source material. DeepSeek reports a 24.9 score on SimpleQA, so a fluent answer alone is insufficient evidence; supply references and ask it to identify unsupported claims.
How should I prompt DeepSeek V3 for JSON?
Explicitly request JSON and include an example of the required structure. Keep each extraction task small enough to return a complete object, and validate the result before using it.
Start using DeepSeek V3 in Instinct
One Spiral account for every tool your agent needs. No subscription; you pay per run.
Paste this into Instinct:
set up https://spiral.ai/SKILL.md


