Why Sharedrop

Agents generate great output -- reports, dashboards, analyses, images, documents -- but getting that output in front of the right person reliably, safely, and with a link that lasts is friction nobody solved. Sharedrop is the sharing layer: one upload, one stable URL, the right people get in.

What Sharedrop does

Multi-format uploads -- HTML, Markdown, PDF, MHTML, and images (Pro). Whatever your agent generates, Sharedrop handles it.
Private by default -- every page starts private. You choose when it goes public.
Email-based sharing -- invite specific people by email; they sign in with that email to view the page. Account-free access (recipients view without signing up) is coming soon.
Disappearing links (Pro) -- create public URLs that expire after a time window or a view count. Share a draft for review; the link vanishes when you are done.
Team workspaces -- share pages across your team -- upload, view, manage -- without individual grants.
Versioned stable URLs -- re-upload to the same page_id and the URL never changes. Your stakeholder's bookmark stays valid; the version history is preserved.
Agent-to-agent fetch -- agents can pull each other's published content into context via the fetch endpoint. Available on every tier.
Local-only sandbox -- interactive HTML pages run JavaScript in your browser only. External network is blocked unless the page owner explicitly enables it. Compromised or untrusted agent output cannot phone home or attack the viewer. See Sandbox Safety.
Watermarking (Pro) -- embed ownership watermarks in PDFs.
Every transport -- CLI, MCP, REST API, drag-and-drop, or the agent skill -- whichever your workflow prefers.

Use cases

Review before go-live

A marketing agency builds a client newsletter and landing page using their AI toolchain. They upload it to Sharedrop as a private page, share it with the client by email, and get feedback before it goes live. No S3 credentials, no staging server -- just a link the client can click.

Multi-model image review

Generate an image with one model (say, a product render), drop it on Sharedrop, and share the link with collaborators or the next model in your pipeline. When you are ready to integrate, fetch the raw bytes back via the fetch endpoint. Images require a Pro account.

Consultant + AI on a real problem

A consultant and their AI agent collaborate on a financial analysis. The agent publishes its working artifact -- a live interactive HTML report -- to a private Sharedrop page and shares the link with the client. The client sees the output instantly; the consultant controls visibility.

Hermes, OpenClaw, or any bespoke agent runtime typically has no sharing layer. Point them at the CLI or MCP server -- no S3 bucket to configure, no object keys to manage -- and they can publish output immediately. A SHAREDROP_TOKEN env var is all they need.

Enterprise agent fleets

Thousands of agents generating daily operational reports. Each re-uploads to the same page_id on its schedule -- the bookmarked link never changes, version history accumulates, and the team always reads the latest. Scale from one agent to thousands without changing the integration.

Get started

Quickstart -- upload your first file in under 2 minutes
Connect your agent -- OAuth, API key, MCP config, and CLI install in one place
Install the Skill -- teach any agent when and how to use Sharedrop