mcp tools for content distribution

MCP stands for Model Context Protocol. it's an open standard — originally from Anthropic, now broadly adopted — that lets AI agents discover and call external tools without custom integration code.

the core idea: an MCP server exposes a set of tools with descriptions, input schemas, and callable functions. an MCP-compatible agent (Claude, Cursor, any LLM with MCP support) connects to the server and can invoke those tools as if they were native capabilities.

from the agent's perspective, it doesn't matter whether a tool is calling a local function or a remote API. MCP abstracts that. the agent sees a tool called create_launch, reads its description, and calls it.

the dropspace MCP server exposes 36 tools across 9 categories. install it with:

npx @jclvsh/dropspace-mcp

set your API key:

export DROPSPACE_API_KEY=ds_live_...

tool categories and what they do:

• launches — create, list, get, update, delete multi-platform launches
• posts — manage individual platform posts within a launch
• personas — create and manage AI writing personas
• media — upload and manage images and videos
• connections — inspect platform OAuth connections
• API keys — create and revoke keys programmatically
• webhooks — register and manage webhook endpoints
• dropspace — product info, supported platforms, version
• usage — check plan usage, remaining launches

here's what an agent workflow looks like with the dropspace MCP server installed. the agent is tasked with distributing a product changelog to social media:

1. agent calls get_dropspace_info to understand which platforms are supported and what content types each accepts
2. agent calls list_personas to find the right brand voice for this product
3. agent calls create_launch with the changelog content, target platforms, and persona ID
4. agent calls publish_launch to trigger publishing
5. agent calls get_launch_status to confirm success and capture post URLs for the report

the agent makes 5 tool calls and the content is live on every connected platform. no custom API integration, no OAuth token management, no per-platform content formatting.

REST APIs work fine for human-written integration code. but for AI agents, tool discovery matters. an agent can't read your API docs — it reads tool descriptions. MCP formalizes that: every tool has a description, input schema, and output schema that the agent uses to decide when and how to call it.

this means an agent that wasn't specifically programmed to use dropspace can still use it — as long as the MCP server is installed and the tool descriptions are clear. that's the composability benefit of MCP: any agent, any framework, same tools.

explore the full docs at dropspace.dev/docs, or join other developers building on dropspace in the community.