OneCLI
Access Control
Open-source credential vault and transparent proxy that lets AI agents call external services without ever seeing API keys.
At a Glance
Pricing
Open Source
Self-hosted open-source credential vault and proxy for AI agents.
Engagement
Available On
CLI
API
Web
SDK
Listed Mar 2026
About OneCLI
OneCLI is an open-source secret vault and transparent MITM proxy designed for AI agents. It stores credentials once in an encrypted vault and injects them into outgoing HTTP requests at runtime, so agents can call external services without ever handling raw API keys. It runs as a single Docker container and requires zero code changes to integrate with any agent framework.
- Encrypted Vault: Credentials are stored with AES-256-GCM encryption and never written to disk in plain text; get started by running
docker run ghcr.io/onecli/onecli. - Transparent Proxy: Set
HTTPS_PROXYto point your agent's HTTP traffic through OneCLI; credentials are injected automatically without modifying agent code. - Web Dashboard: Manage agents, secrets, and permissions from a local web UI at
localhost:10254. - Full Audit Trail: Every API call is logged with agent identity and timestamp, enabling real-time approval or denial of actions.
- Human-in-the-Loop Policies: Define per-action policies so a human can approve or deny agent actions before they execute.
- One-Command Setup: The entire proxy, vault, and dashboard stack runs from a single
docker runcommand. - Framework Agnostic: Works with any agent framework including OpenClaw, NanoClaw, IronClaw, Dify, n8n, OpenHands, and more.
- Revocation: Revoke credentials once from the vault and access is removed everywhere instantly.
- Language SDKs: Programmatic integration available via language SDKs for developers who need deeper control.
Community Discussions
Be the first to start a conversation about OneCLI
Share your experience with OneCLI, ask questions, or help others learn from your insights.
Pricing
OPEN SOURCE
Open Source
Self-hosted open-source credential vault and proxy for AI agents.
- Encrypted credential vault (AES-256-GCM)
- Transparent MITM proxy
- Web dashboard
- Full audit trail
- Human-in-the-loop policies
Capabilities
Key Features
- Encrypted credential vault (AES-256-GCM)
- Transparent MITM proxy for credential injection
- Zero code changes required
- Web dashboard for managing agents and secrets
- Full audit trail of all API calls
- Human-in-the-loop approval policies
- Single Docker container deployment
- Per-action permission policies
- Real-time approve/deny controls
- Language SDKs for programmatic integration
Integrations
OpenClaw
NanoClaw
IronClaw
Dify
n8n
OpenHands
Google
GitHub
Slack
Gmail
API Available
View Docs