AI Gateway + Assistant

Route AI safely across assistants, agents, and models

Skitflow brings AI Gateway and AI Assistant into the deployment plane: route models through BYOK or platform providers, let teams use external agents through gateway APIs, and turn operational work into explicit proposals instead of mystery mutations.

AI control plane

Gateway, Assistant, and governed operations

Centralize model access and give teams one governed path for Skitflow Assistant, custom agents, API keys, and runtime-aware operations.

AI Gateway

Route Assistant, Ask AI, OpenClaw, Hermes Agent, or your own agents through tenant BYOK providers or platform-managed models.

AI Assistant

Generate Compose, review and harden configs, inspect deployments, and prepare confirmed operations from inside Skitflow.

Agent API access

Use AI Gateway API keys so external agents can call approved models without copying provider secrets outside the platform.

MCP server integration

AI agents can deploy and manage services through Skitflow MCP without custom glue code.

Safe proposals

Mutating work is proposal-based: users confirm first, then Skitflow re-checks authorization server-side.

Observability context

When observability is enabled, the Assistant can use logs, metrics, route data, and deployment history for diagnostics.

Bring your own agents

Use OpenClaw, Hermes Agent, or your own agent stack

Skitflow does not force one agent runtime. Keep model access, provider governance, and deployment actions inside Skitflow while your preferred agents call the gateway and MCP APIs.

Bring governed AI operations to the whole team

Give everyone a clear path from prompt to production, with AI Gateway routing, external agent access, and Assistant proposals keeping control in the platform.