How to Use Particle-1 in Instinct
Particle-1 is built for debugging, code changes, and data engineering. Post-trained on DeepSeek V4.1 Flash, it targets fewer reasoning and tool-call tokens while keeping task accuracy close to that model.
Prompts to try with Particle-1 in Instinct
Copy an example into Instinct, or ask in your own words.
Each morning, use Particle-1 to review the supplied build-error feed. Group repeated failures and give each distinct error a likely cause and next check.
After each failed data job, use Particle-1 to inspect the supplied SQL and error messages. Flag type mismatches and propose targeted validation queries.
Every weekday, use Particle-1 to review incoming bug descriptions. Draft reproduction steps and acceptance criteria, and mark missing evidence.
Each evening, use Particle-1 to compare today's supplied test failures with yesterday's. Identify new failures and propose one investigation step for each.
Every Friday, use Particle-1 to review the supplied engineering change summaries. List changes that need explicit edge-case tests or documentation.
How to set up Particle-1 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 Particle-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": "particle-ai/particle-1", "input": { "messages": [ { "role": "system" } ] } }'
Then paste one of the prompts above: Instinct finds Particle-1 on Spiral, shows the price and runs it.
Particle-1 providers and prices
Spiral runs Particle-1 through OpenRouter. Open a card for today's price, with every fee included.
Up to $0.002667 per run at the default inputs. $0.091 per 1M input tokens, $0.65 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 | ||
| Logprobs | on or off | ||
| Max tokens | 1 to 131072 | The most output tokens to generate. | |
| Temperature | 0 to 2 | ||
| Top logprobs | 0 to 20 | ||
| Presence penalty | -2 to 2 | ||
| Reasoning effort | none, minimal, low, medium, high, xhigh, or max | high | How much the model reasons. |
| Frequency penalty | -2 to 2 | ||
| Repetition penalty | 0 to 2 |
Every API field of Particle AI: Particle-1 is on its tool page: open its card above.
What can you do with Particle-1 in Instinct?
Turn recurring engineering errors into investigation briefs with Particle-1. Use scheduled reviews to collect evidence, rank likely causes, and propose checks for an engineer.
Fix a failing build
Use Particle-1 to inspect compiler errors, identify the cause, propose a targeted fix, and verify the build.
Repair a React interaction
Ask Particle-1 to correct broken navigation or row selection and check the affected behavior.
Correct dbt transformations
Have Particle-1 trace inconsistent types or key generation through a dbt project and propose tests.
Implement a codebase change
Give Particle-1 a concrete requirement and acceptance criteria, then request an implementation with regression checks.
Plan a debugging session
Provide an error trace and relevant code. Ask Particle-1 for ranked hypotheses and checks that distinguish them.
Triage recurring engineering failures
Use Particle-1 to review incoming build errors or data-job failures and draft investigation steps.
Produce structured engineering reports
Ask Particle-1 for JSON summaries of errors, proposed fixes, and verification evidence, then validate the results.
More on Particle-1 and Instinct
Overview
Particle-1 in other agents
More Instinct guides
Questions about Particle-1 in Instinct
Can Instinct use Particle-1?
Yes. Add Spiral to Instinct as an MCP server, and Instinct can find and run Particle-1 like any built-in tool.
How much does Particle-1 cost in Instinct?
Up to $0.002667 per run at the default inputs. $0.078 per 1M input tokens, $0.65 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 Particle-1 in Instinct?
You need no OpenRouter account. Instinct uses one Spiral API key for every tool, and Spiral runs Particle-1 on its own OpenRouter account.
How long does it take to set up Particle-1 in Instinct?
About two minutes: create an API key, then send Instinct this message, then give Instinct the key.
Does Instinct wait while Particle-1 runs?
Particle-1 runs in the background. Instinct starts the run, checks on it and gives you the text, json when it's ready.
What is Particle-1 best suited for?
Particle-1 focuses on software engineering and data engineering. Its published evaluations cover terminal tasks, React fixes, dbt projects, and codebase changes.
How does Particle-1 compare with DeepSeek V4.1 Flash?
Particle-1 is post-trained on DeepSeek V4.1 Flash to use fewer tokens. Particle reports 28.4% to 52.2% fewer output tokens across four engineering benchmarks, with pass rates within the base model's confidence intervals.
How much context and output can Particle-1 handle?
The context window holds 1,048,576 tokens across the prompt and completion. The output ceiling is 131,072 tokens, including reasoning.
When should I reduce Particle-1's reasoning?
Consider lower effort for straightforward explanations or narrow edits, and compare results on representative tasks. Reasoning is enabled by default and consumes part of the output allowance.
Can I rely on Particle-1 to match a JSON schema?
Validate the result before your application uses it. Particle's documentation distinguishes valid JSON from guaranteed compliance with your application's schema.
Start using Particle-1 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


