What does Sociable AI do?

Sociable AI helps social, community, and growth teams discover relevant social posts and organize timely, on-brand engagement with human review.

Does Sociable AI have an API?

Yes. The authenticated Search API supports natural-language Instagram content search and structured post-detail retrieval through REST and MCP.

Who is Sociable AI for?

It is for brands, agencies, and social teams that need repeatable content discovery, community engagement, analytics, or social listening workflows.

Official developer resources

Build with the Sociable AI Search API

Search real-time social content with natural language and retrieve structured post detail. The API is designed for server-to-server use by products, analysts, and agents. It uses bearer API keys, JSON requests and responses, idempotent search retries, stable error codes, and documented rate-limit headers.

Resources

Search API documentation

Authentication, natural-language social-content search, post detail, retries, quotas, and rate limits.

OpenAPI specification

Machine-readable operations, parameters, authentication, request bodies, responses, and error schemas.

Agent instructions

What Sociable AI does, when an agent should use it, limitations, and canonical resource links.

API catalog

A machine-readable catalog that points software agents to the public Search API description.

MCP server

Authenticated Streamable HTTP tools and resources for MCP-compatible agents and developer clients.

Agent Skill

A digest-verified skill with search workflow, retry behavior, credential safety, and human-review guidance.

Documentation MCP

A separate read-only MCP server for finding and reading official product, pricing, authentication, and API documentation.

NLWeb ask endpoint

Natural-language discovery over canonical public resources with Schema.org JSON or server-sent events.

Quick start

Request a Search API key from Sociable AI, then send it as a bearer token. Search requests should include an Idempotency-Key so a safe retry cannot create a duplicate search. Use a client timeout of at least 270 seconds for long-running searches.

curl -X POST \
  https://search-backend-s4c56s44ia-uk.a.run.app/v1/search \
  -H "Authorization: Bearer $SOCIABLE_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: unique-request-id" \
  -d '{"query":"funny cooking videos from the last 30 days"}'

Ask the documentation

Search canonical product and developer resources through the public NLWeb endpoint. The same form is exposed as a read-only declarative WebMCP tool in supported browsers.