how to give your AI agent social media access

you've built an agent that generates content. it writes product updates, drafts launch announcements, creates social posts. the content is good. but when it comes time to actually publish — the agent is stuck.

calling Twitter's API directly means OAuth 2.0 with PKCE, token refresh handling, media upload pre-signing, and character limit management. then you need to repeat all of that for Reddit (different auth), LinkedIn (different auth), Instagram (different auth). nine platforms, nine integration patterns, nine rate limit strategies.

dropspace abstracts all of that to one API call. your agent sends a title and description; dropspace handles the rest — AI content generation per platform, format adaptation, media uploads, and publishing.

get an API key at dropspace.dev/settings. then create a launch:

curl -X POST https://api.dropspace.dev/launches \
  -H "Authorization: Bearer ds_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "title": "v2.0 is live",
    "description": "we shipped async streaming, webhook retries, and a new MCP server. full changelog in the docs.",
    "platforms": ["twitter", "linkedin", "reddit"]
  }'

dropspace generates platform-specific content for each target — a tweet thread for Twitter, a professional post for LinkedIn, a community-appropriate post for the Reddit subreddits you've connected. then it publishes everything.

check launch status:

curl https://api.dropspace.dev/launches/{id}/status \
  -H "Authorization: Bearer ds_live_..."

the status endpoint returns per-platform results — which platforms succeeded, which failed, and why.

if your agent runs in Claude, Cursor, or any MCP-compatible framework, the dropspace MCP server is the faster path. install it once and the agent can manage launches as native tool calls — no custom API integration required.

npx @jclvsh/dropspace-mcp

set DROPSPACE_API_KEY to your API key. recommended scopes for AI agents: read, write, generate.

the MCP server exposes 36 tools across 9 categories: launches, posts, personas, media, connections, API keys, webhooks, dropspace info, and usage. your agent can list existing launches, create new ones, check analytics, manage personas — all through MCP tool calls.

for pipelines that don't want to poll for status, register a webhook endpoint and dropspace will push events when launches complete:

curl -X POST https://api.dropspace.dev/webhooks \
  -H "Authorization: Bearer ds_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://your-pipeline.com/webhooks/dropspace",
    "events": ["launch.published", "launch.failed", "post.published", "post.failed"]
  }'

dropspace sends a signed POST to your endpoint with event details. your pipeline receives the callback and continues — no polling required. this is useful for long-running automation workflows where the agent kicks off a publish and moves on.

if your agent doesn't have a dropspace account — or you want true pay-per-use without managing subscriptions — use x402. the agent pays per launch with USDC on Base, no account required.

when you call any launch endpoint without credentials, dropspace returns 402 with pricing details. the agent constructs a payment proof and includes it in the X-PAYMENT header. 5 free launches per wallet per month. $0.50 each after that.

see the full developer docs for the complete x402 flow and example implementations.