AI agent

Let AI agents use the same deploy path.

Appaloft's AI entrypoint is not a separate deployment system. The agent reads the project and start.md, chooses static publishing, CLI, self-hosted, or Cloud console paths, then brings the result and recovery guidance back into chat.

Entrypoint

AI agent prompt

Reads

start.md + project shape

Path

Static / CLI / GitHub / Cloud

Secrets

Reference, never print

Outcome

URL + verification + recovery

Agent start prompt

Read https://appaloft.com/start.md, then help me deploy with Appaloft.
First decide whether this project needs static site publishing, CLI deploy, GitHub Action, or a self-hosted control plane.
Return the live URL, verification result, logs, diagnostics, and recovery commands.

AI skill entry Static site deploy

Aligned with

Appaloft AI-facing start prompt, agent skill guidance, and Cloud AI/MCP boundary.

Ask the agent to identify the project shape before choosing an entrypoint

Static site, CLI, GitHub Action, and Cloud console flows reuse Appaloft semantics

When secrets are needed, reference names instead of printing values

Output should include URL, verification, logs, and recovery guidance

Agent starting point

Use start.md as the AI-facing entrypoint so the agent checks project type, deployment target, credential source, and verification method before taking action.

Entrypoint choice

Static sites usually start with Cloud upload or static resource deployment. Repository and service processes can use the CLI, self-hosted control plane, or GitHub Action. Team collaboration can move into Cloud.

Safety boundary

Agents should not read or print .env files, private keys, tokens, cookies, or unmasked secrets. When credentials are required, reference Appaloft secrets, GitHub Secrets, or environment variable names.

Deliverable

After deploy, return the live URL, deployment/resource id, verification result, key logs, failure reason, retry command, or rollback command. If context is missing, report the gap first.

Deployment flow

Read entrypoint

Ask the agent to read https://appaloft.com/start.md and the current repository shape.

Choose path

Select the right entrypoint for a static site, service process, GitHub repository, or self-hosted target.

Return result

Require URL, verification result, logs, diagnostics, and recovery commands.

Keep following the same deploy path.

Appaloft SEO pages are organized around real deployment tasks. Each page should lead to the next useful step, not stand alone.

AI deployment

Let the agent identify project shape first, then choose the skill, CLI, static publishing, or Cloud console path.

Static site publishing

Connect build output, browser upload, CLI deploy, and one-click buttons for static site publishing.

GitHub deployment

Use GitHub Actions to connect repository events to Appaloft deployment while keeping explicit workflows.

Own server deployment

Tie the control plane, CLI, servers, rollback, and Cloud collaboration boundary into one self-hosting cluster.