connect · five clients

Drop DropFast into your agent.

DropFast publishes over a plain REST API. For each client below, run one command to install the prebuilt dropfast skill and the @dropfast/cli, then dropfast login — you're wired up. A hosted MCP server is also live at dropfast.dev/api/mcp for clients that speak remote MCP — connect it from Claude, ChatGPT, or your IDE.

Primary path: install the prebuilt skill + @dropfast/cli in one command per client (it drops the skill into that client's own skills dir), then run dropfast login. No npm or offline? Each card also carries a fallback self-setup prompt.

Working in a terminal? The first-party dropfast CLI wraps the whole API — publish, search, bulk permission changes, and a machine-readable command tree via dropfast llms.

Don't have a key yet? Generate one →

Claude Code

claude

Anthropic's coding agent.

Paste to install the skill + CLI:

bash

npm install -g @dropfast/cli && \
  mkdir -p ~/.claude/skills/dropfast && \
  curl -fsSL https://dropfast.dev/dropfast.SKILL.md -o ~/.claude/skills/dropfast/SKILL.md && \
  dropfast login

Fallback — no npm / offline (agent writes its own skill)

Paste into the chat and the agent reads the REST docs and writes its own DropFast skill — no install required:

text

Set yourself up to publish HTML artifacts to DropFast via its REST API, so you can hand me stable URLs instead of pasting code into chat.

1. Read https://dropfast.dev/llms.txt to learn the API.
2. Ask me for a DropFast API key — I'll create one at dropfast.dev/dashboard/api-keys. Store it as a secret (env var / secret store), never in a committed file.
3. Write yourself a .claude/skills/dropfast/SKILL.md (with name + description frontmatter) capturing the workflow so it persists across sessions:
   - Publish: POST https://dropfast.dev/api/v1/sites (multipart file=index.html, name=...) -> returns the URL. Echo the URL in your reply.
   - Update in place: PUT /api/v1/sites/{slug} so the URL stays stable. One slug per artifact, not per turn.
   - Default accessMode "private"; flip to "public"/"password" only when I ask.
   - Inline review: publish with commentsEnabled=true, then GET /api/v1/sites/{slug}/comments and resolve each with PATCH /api/v1/comments/{id} after you address it.
   - Auth every request with: Authorization: Bearer df_sk_...
4. Confirm by publishing a one-line test page and showing me the URL.

Then ask me exactly one question, and remember my answer:

  When should I publish artifacts to DropFast?
    1. Every HTML/Markdown artifact I generate.
    2. Only shareable deliverables — plans, prototypes, reports, reviews.
    3. Only when I explicitly ask.

Record the chosen mode in the skill file so the habit persists across sessions.

Cursor

cursor

AI-first editor.

Paste to install the skill + CLI:

bash

npm install -g @dropfast/cli && \
  mkdir -p ~/.cursor/skills/dropfast && \
  curl -fsSL https://dropfast.dev/dropfast.SKILL.md -o ~/.cursor/skills/dropfast/SKILL.md && \
  dropfast login

Fallback — no npm / offline (agent writes its own skill)

Paste into the chat and the agent reads the REST docs and writes its own DropFast skill — no install required:

text

Set yourself up to publish HTML artifacts to DropFast via its REST API, so you can hand me stable URLs instead of pasting code into chat.

1. Read https://dropfast.dev/llms.txt to learn the API.
2. Ask me for a DropFast API key — I'll create one at dropfast.dev/dashboard/api-keys. Store it as a secret (env var / secret store), never in a committed file.
3. Write yourself a rule file under .cursor/rules/ capturing the workflow so it persists across sessions:
   - Publish: POST https://dropfast.dev/api/v1/sites (multipart file=index.html, name=...) -> returns the URL. Echo the URL in your reply.
   - Update in place: PUT /api/v1/sites/{slug} so the URL stays stable. One slug per artifact, not per turn.
   - Default accessMode "private"; flip to "public"/"password" only when I ask.
   - Inline review: publish with commentsEnabled=true, then GET /api/v1/sites/{slug}/comments and resolve each with PATCH /api/v1/comments/{id} after you address it.
   - Auth every request with: Authorization: Bearer df_sk_...
4. Confirm by publishing a one-line test page and showing me the URL.

Then ask me exactly one question, and remember my answer:

  When should I publish artifacts to DropFast?
    1. Every HTML/Markdown artifact I generate.
    2. Only shareable deliverables — plans, prototypes, reports, reviews.
    3. Only when I explicitly ask.

Record the chosen mode in the skill file so the habit persists across sessions.

Codex

codex

OpenAI Codex CLI.

Paste to install the skill + CLI:

bash

npm install -g @dropfast/cli && \
  mkdir -p ~/.codex/skills/dropfast && \
  curl -fsSL https://dropfast.dev/dropfast.SKILL.md -o ~/.codex/skills/dropfast/SKILL.md && \
  dropfast login

Fallback — no npm / offline (agent writes its own skill)

Paste into the chat and the agent reads the REST docs and writes its own DropFast skill — no install required:

text

Set yourself up to publish HTML artifacts to DropFast via its REST API, so you can hand me stable URLs instead of pasting code into chat.

1. Read https://dropfast.dev/llms.txt to learn the API.
2. Ask me for a DropFast API key — I'll create one at dropfast.dev/dashboard/api-keys. Store it as a secret (env var / secret store), never in a committed file.
3. Write yourself your AGENTS.md capturing the workflow so it persists across sessions:
   - Publish: POST https://dropfast.dev/api/v1/sites (multipart file=index.html, name=...) -> returns the URL. Echo the URL in your reply.
   - Update in place: PUT /api/v1/sites/{slug} so the URL stays stable. One slug per artifact, not per turn.
   - Default accessMode "private"; flip to "public"/"password" only when I ask.
   - Inline review: publish with commentsEnabled=true, then GET /api/v1/sites/{slug}/comments and resolve each with PATCH /api/v1/comments/{id} after you address it.
   - Auth every request with: Authorization: Bearer df_sk_...
4. Confirm by publishing a one-line test page and showing me the URL.

Then ask me exactly one question, and remember my answer:

  When should I publish artifacts to DropFast?
    1. Every HTML/Markdown artifact I generate.
    2. Only shareable deliverables — plans, prototypes, reports, reviews.
    3. Only when I explicitly ask.

Record the chosen mode in the skill file so the habit persists across sessions.

Gemini CLI

gemini

Google's gemini-cli.

Paste to install the skill + CLI:

bash

npm install -g @dropfast/cli && \
  mkdir -p ~/.gemini/skills/dropfast && \
  curl -fsSL https://dropfast.dev/dropfast.SKILL.md -o ~/.gemini/skills/dropfast/SKILL.md && \
  dropfast login

Fallback — no npm / offline (agent writes its own skill)

Paste into the chat and the agent reads the REST docs and writes its own DropFast skill — no install required:

text

Set yourself up to publish HTML artifacts to DropFast via its REST API, so you can hand me stable URLs instead of pasting code into chat.

1. Read https://dropfast.dev/llms.txt to learn the API.
2. Ask me for a DropFast API key — I'll create one at dropfast.dev/dashboard/api-keys. Store it as a secret (env var / secret store), never in a committed file.
3. Write yourself your AGENTS.md capturing the workflow so it persists across sessions:
   - Publish: POST https://dropfast.dev/api/v1/sites (multipart file=index.html, name=...) -> returns the URL. Echo the URL in your reply.
   - Update in place: PUT /api/v1/sites/{slug} so the URL stays stable. One slug per artifact, not per turn.
   - Default accessMode "private"; flip to "public"/"password" only when I ask.
   - Inline review: publish with commentsEnabled=true, then GET /api/v1/sites/{slug}/comments and resolve each with PATCH /api/v1/comments/{id} after you address it.
   - Auth every request with: Authorization: Bearer df_sk_...
4. Confirm by publishing a one-line test page and showing me the URL.

Then ask me exactly one question, and remember my answer:

  When should I publish artifacts to DropFast?
    1. Every HTML/Markdown artifact I generate.
    2. Only shareable deliverables — plans, prototypes, reports, reviews.
    3. Only when I explicitly ask.

Record the chosen mode in the skill file so the habit persists across sessions.

JetBrains Junie

junie

JetBrains AI assistant.

Paste to install the skill + CLI:

bash

npm install -g @dropfast/cli && \
  mkdir -p ~/Library/Application Support/JetBrains/Junie/skills/dropfast && \
  curl -fsSL https://dropfast.dev/dropfast.SKILL.md -o ~/Library/Application Support/JetBrains/Junie/skills/dropfast/SKILL.md && \
  dropfast login

Fallback — no npm / offline (agent writes its own skill)

Paste into the chat and the agent reads the REST docs and writes its own DropFast skill — no install required:

text

Set yourself up to publish HTML artifacts to DropFast via its REST API, so you can hand me stable URLs instead of pasting code into chat.

1. Read https://dropfast.dev/llms.txt to learn the API.
2. Ask me for a DropFast API key — I'll create one at dropfast.dev/dashboard/api-keys. Store it as a secret (env var / secret store), never in a committed file.
3. Write yourself your project instructions file capturing the workflow so it persists across sessions:
   - Publish: POST https://dropfast.dev/api/v1/sites (multipart file=index.html, name=...) -> returns the URL. Echo the URL in your reply.
   - Update in place: PUT /api/v1/sites/{slug} so the URL stays stable. One slug per artifact, not per turn.
   - Default accessMode "private"; flip to "public"/"password" only when I ask.
   - Inline review: publish with commentsEnabled=true, then GET /api/v1/sites/{slug}/comments and resolve each with PATCH /api/v1/comments/{id} after you address it.
   - Auth every request with: Authorization: Bearer df_sk_...
4. Confirm by publishing a one-line test page and showing me the URL.

Then ask me exactly one question, and remember my answer:

  When should I publish artifacts to DropFast?
    1. Every HTML/Markdown artifact I generate.
    2. Only shareable deliverables — plans, prototypes, reports, reviews.
    3. Only when I explicitly ask.

Record the chosen mode in the skill file so the habit persists across sessions.