> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cynative.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Cynative Quickstart: Install, Configure, and Run in Minutes
> Install Cynative, configure an LLM provider, and run your first cloud security audit with a built-in agent or a natural-language question.
Cynative is a single static binary with no dependencies. Install it, point it at an LLM provider and your cloud credentials, and you're ready to run security research against live infrastructure — all read-only, all verified.
The fastest way to install on macOS or Linux is Homebrew:
```bash theme={null}
brew install cynative/tap/cynative
```
**Install script** (macOS / Linux) — downloads the latest release, verifies the SHA-256 checksum against the release `checksums.txt`, and installs to `~/.local/bin` by default:
```bash theme={null}
curl -fsSL https://raw.githubusercontent.com/cynative/cynative/main/install.sh | sh
```
**Windows** (Scoop):
```powershell theme={null}
scoop bucket add cynative https://github.com/cynative/scoop-bucket
scoop install cynative
```
See the [Installation guide](/installation) for manual downloads, release signature verification, version pinning, and macOS `.pkg` details.
Cynative requires an LLM provider and model. Set them via environment variables before your first run. The example below uses Anthropic:
```bash theme={null}
export CYNATIVE_LLM_PROVIDER=anthropic
export CYNATIVE_LLM_MODEL=claude-opus-5-5
export ANTHROPIC_API_KEY=...
```
Cynative supports almost any AI provider through the embedded [Bifrost](https://github.com/maximhq/bifrost) SDK, including OpenAI, Azure OpenAI, Amazon Bedrock, Google Vertex/Gemini, Groq, Ollama, vLLM, and more. A few quick alternatives:
```bash theme={null}
# OpenAI
export CYNATIVE_LLM_PROVIDER=openai
export CYNATIVE_LLM_MODEL=gpt-5.6-sol
export OPENAI_API_KEY=sk-...
# Amazon Bedrock (uses your existing AWS credential chain)
export CYNATIVE_LLM_PROVIDER=bedrock
export CYNATIVE_LLM_MODEL=anthropic.claude-opus-5
export CYNATIVE_LLM_BEDROCK_REGION=us-east-1
# Google Vertex
export CYNATIVE_LLM_PROVIDER=vertex
export CYNATIVE_LLM_MODEL=gemini-3.1-pro-preview
export CYNATIVE_LLM_VERTEX_PROJECT_ID=my-gcp-project
export CYNATIVE_LLM_VERTEX_REGION=global
```
Provider settings can also be written to `~/.cynative/config.yaml` for persistence. See the providers documentation for every supported provider's configuration reference.
Run `cynative doctor` to validate your configuration and connector readiness without starting a research session:
```bash theme={null}
cynative doctor
```
Add `--live-llm` to also probe the configured model with a tool-less round-trip, confirming your API key and endpoint are reachable:
```bash theme={null}
cynative doctor --live-llm
```
Fix any issues reported before continuing — `doctor` checks your LLM configuration, credential discovery for each connector, and action-gate policy resolution.
Cynative ships 45 embedded agents for AWS, GCP, Azure, GitHub, and Kubernetes. Run one non-interactively with the `-p` flag:
```bash theme={null}
cynative -p --agent aws-network-exposure
```
The `-p` / `--print` flag runs a single task non-interactively and exits — ideal for scripts and CI. Without `-p`, Cynative opens an interactive session after the agent finishes.
Each tool call waits for a single keystroke approval: `y` runs it once, `a` approves all later calls to that tool for the session, and any other key denies. Use `--auto-approve` to skip approvals for unattended runs:
```bash theme={null}
cynative -p --auto-approve --agent aws-network-exposure
```
To browse the full catalog:
```bash theme={null}
cynative agents list # every agent with its description
cynative agents show aws-network-exposure # the exact prompt that would run
```
You can ask Cynative anything in natural language. It picks up the credentials already in your shell and reasons through your live infrastructure:
```bash theme={null}
cynative -p "what is publicly exposed that shouldn't be?"
```
More examples:
```bash theme={null}
cynative -p "which IAM roles can escalate to admin?"
cynative -p "high-risk cloud permissions, trace each to the PR where it was granted"
cynative -p "cloud credentials leaked in source code and their current blast radius"
cynative "live cloud resources absent from IaC - drift" # opens an interactive session
cat findings.json | cynative -p "triage these findings by exploitability"
```
Cynative prints a short footer with timing and token usage to stderr, so redirecting stdout keeps the captured answer clean:
```bash theme={null}
cynative -p "what is publicly exposed that shouldn't be?" > findings.txt
```
## Next Steps
Browse all 45 built-in agents for AWS, GCP, Azure, GitHub, and Kubernetes — with one-line descriptions and example use cases.
An agent is a single markdown file. Copy a built-in, edit the prompt, and run it under a new name — no code required.
Learn how Cynative discovers credentials, enforces read-only at three layers (network, action gate, and credential), and connects to each cloud and platform.
All configuration keys and environment variables — LLM provider settings, resource and cost controls, audit log options, and connector hardening.
This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.