# A swarm of hackers in your terminal.

> Excession Labs builds Incursion, a terminal AI agent for authorized security testing, and provides model APIs. You tell it which system to test, and it splits the work into tasks that several hackers run at the same time. Each finding includes the evidence, the steps to reproduce it, and a suggested fix.

Source page: https://excessionlabs.ai/ · Company: Excession Labs · Contact: support@excessionlabs.ai

For pentesters, red teams and bug bounty hunters doing authorized work.

Install Incursion in one line (macOS, Linux, Windows):

```sh
# macOS and Linux
curl -fsSL https://excessionlabs.ai/install.sh | sh

# Windows (PowerShell)
irm https://excessionlabs.ai/install.ps1 | iex
```

1. Run `excession login` and approve in your browser.
2. Pick a plan, or start with your $5 free signup credit.
3. Run `excession` and say which system to test.

[Start free with $5 credit](https://excessionlabs.ai/app). New accounts get $5 of credit once, with no card. [Pricing](https://excessionlabs.ai/pricing)

## A scripted security review

- **Incursion plans the work.** It splits your request into tasks.
- **Several hackers run them at once.** Each tool runs on its own, waits for your approval, or is blocked, as you choose.
- **Every finding has evidence.** A task is done only when each finding links to its evidence.

On phones, select “Show the terminal” to open the recording.

1. **The brief.** Ask for a review of sample API configuration and audit logs.
2. **Objectives set.** The coordinator sets three objectives: logging, audit coverage, and release controls.
3. **Three hackers, in parallel.** Open /agents to see each assignment, its status, and the hacker's latest action.
4. **Inspect the evidence.** Open a hacker to read the file checks behind its conclusions.
5. **Review the risks.** The run records two configuration risks, each with a severity and its evidence.
6. **Prioritize the fixes.** The final report explains what to change and which control checks passed.
7. **All three hackers finished.** Each hacker's notes and findings stay available.
8. **Revisit the results.** Use /results to return to the recorded risks after the review.

### Sample finding

Request bodies are logged without field redaction (Medium). The sample API configuration enables body logging without sensitive-field redaction. Disable request-body logging and define a redaction allowlist. The finding remains pending until retested.

## One plan for Incursion and your coding agent

Plan credit pays for Incursion and for API calls from Claude Code, OpenCode, Codex or your own code. Every plan includes:

- Incursion, the security agent for your terminal
- API keys for Claude Code, OpenCode, Codex and your own code
- DeepSeek V4.1 Flash Abliterated with a 1M-token context
- New credit every month, for one person

Table 1. Plans in USD, billed monthly. Plan credit is used before your balance and expires at renewal.

| Plan | Price | Plan credit |
| --- | --- | --- |
| Free | $0 | $5, once, when you sign up. No card |
| Starter | $20 | $30 (+50%). Side projects and CTFs |
| Pro (recommended) | $100 | $175 (+75%). Daily work for one person |
| Max | $200 | $400 (+100%). Long runs and large codebases |
| Ultra | $500 | $1,250 (+150%). Many agents in parallel |

Rather not subscribe? Add funds from $0.50. Your balance never expires.

Table 2. USD per 1M tokens. The Cached column prices cached input.

| Model | Input | Cached | Output |
| --- | --- | --- | --- |
| DeepSeek V4.1 Flash Abliterated | $0.20 | $0.025 | $0.80 |

More on plans, balance and limits: https://excessionlabs.ai/pricing.md

## No refusals on authorized work

Incursion takes on the offensive work you are authorized to do, without refusing it: recon, web and API testing, code audit, exploit development, reverse engineering and malware analysis. The [Acceptable Use Policy](https://excessionlabs.ai/aup/) still blocks weapons, CBRN and CSAM.

## Questions

### Who is it for?

Pentesters, red teams, bug bounty hunters and security engineers doing authorized work.

### Which model does it use?

DeepSeek V4.1 Flash Abliterated, with a 1M-token context. It runs on your Excession account, so you need no other provider keys.

### Can I use my plan in Claude Code, OpenCode or Codex?

Yes. Create an API key in the console and set it in any tool that accepts an OpenAI or Anthropic base URL. Incursion and your API keys spend the same credit. Setup guides: [Claude Code](https://excessionlabs.ai/blog/deepseek-claude-code/), [OpenCode](https://excessionlabs.ai/blog/deepseek-opencode/).

### How do I pay?

Choose a monthly plan, or add $0.50 to $1,000 at a time to your balance. You pay in US dollars, by card or crypto. The console shows every request and its cost.

### Does it work with my tools?

Yes. MCP servers, skills, `AGENTS.md`, hooks, headless runs for CI, and saved sessions are covered in the [docs](https://excessionlabs.ai/docs/).

## Install Incursion

Then run `excession login`, pick a plan or use your $5 free signup credit, and run `excession`.

[Choose a plan](https://excessionlabs.ai/app/billing)

## Building your own tools?

The Excession API serves the same models on the same account. [Explore the API](https://excessionlabs.ai/api)

## More

- Docs: https://excessionlabs.ai/docs.md
- API: https://excessionlabs.ai/api.md
- Blog: https://excessionlabs.ai/blog.md
- All site text in one file: https://excessionlabs.ai/llms-full.txt