Registry Copy Pack

Pharlo — the content-posting MCP server and API that lets AI agents publish, schedule, and track social media.

Pharlo is an MCP server and REST API that lets AI agents and apps publish, schedule, and track social media content. Connect channels via OAuth, publish to YouTube and Facebook, get signed webhooks on delivery, and read analytics — one surface across platforms.

Pharlo is a content-posting MCP server and REST API for AI agents and developers. Through 21 MCP tools (or the equivalent REST endpoints) you can connect social channels via OAuth, publish and schedule posts, retry failed deliveries, and read analytics — without integrating each platform's API yourself. Live platforms today are YouTube and Facebook; Instagram and TikTok are planned. Auth is OAuth 2.0 (PKCE) for MCP and API keys for REST, deliveries are HMAC-signed, and usage is credit-metered.

Pharlo is a content-posting service exposed as both an MCP server and a REST API, built for AI agents and developers rather than humans clicking a UI.

Connect a social channel once via OAuth and Pharlo manages the platform tokens for you. From there, an agent can publish immediately, schedule for later, cancel or update scheduled posts, retry failed deliveries, and read analytics — all through 21 typed MCP tools or the equivalent REST endpoints. Everything is discoverable: a machine-readable tools catalog, an OpenAPI 3.0 spec, and a .well-known manifest.

Live platforms today are YouTube and Facebook; Instagram and TikTok are planned. MCP auth is OAuth 2.0 with PKCE; REST uses API keys. Webhook deliveries are signed with HMAC-SHA256 so callers can verify authenticity, and usage is metered in credits. Claude is tested; Cursor, Windsurf, ChatGPT, and n8n are supported over remote OAuth MCP.

Pharlo is a content-posting platform for AI agents and developers, exposed through two surfaces that share one backend: a remote MCP server (Streamable HTTP, OAuth 2.0) and a REST API (API-key auth, OpenAPI 3.0).

The core idea is integrate once, publish anywhere we support. Instead of wiring up each social platform's API — with its own auth, schemas, rate limits, and quirks — you connect a channel to Pharlo via OAuth, and Pharlo manages the platform tokens server-side. An agent then works against a single, consistent surface of 21 typed MCP tools (or matching REST endpoints): list platforms and capabilities, connect channels, create assignments to publish now or on a schedule, cancel or update scheduled posts, update already-published content, retry failed deliveries, and pull analytics and post stats.

Pharlo is built to be machine-discoverable. There's a machine-readable MCP tools catalog (names, input schemas, behavior annotations, per-tool credit costs, and worked examples), an OpenAPI 3.0 specification, a .well-known MCP manifest, and an error catalog mapping each failure to the recovery action an agent should take. Webhook deliveries are signed with HMAC-SHA256 (constant-time verification) so callers can trust delivery events. Usage is metered in credits, with clear HTTP 402 codes when credits run out or a spending limit is hit.

Live platforms today are YouTube and Facebook. Instagram and TikTok are planned (modelled internally but not yet exposed for publishing); Twitter / X is under exploration. Pharlo never silently drops work — failed publishes surface a reason and can be retried explicitly.

Authentication is OAuth 2.0 with PKCE for the MCP server and API keys for REST. Pharlo is multi-tenant with workspace isolation. Claude is tested as an MCP client; Cursor, Windsurf, ChatGPT, and n8n are supported over remote OAuth MCP. Pharlo is a hosted, remote service — there is no self-hosted build, and the core is closed-source, though example configs, REST snippets, webhook verifiers, and n8n templates are published for reuse.

- [Pharlo](https://pharlo.io/_mcp) — Publish, schedule, and track social media (YouTube, Facebook) on a user's behalf. Remote Streamable HTTP, OAuth 2.0, 21 tools, signed webhooks.

Let your AI agent publish, schedule, and track social posts via MCP

Show HN: Pharlo – an MCP server and API for AI agents to publish to social media

Pharlo: an MCP server + REST API that lets AI agents publish and schedule social posts (YouTube, Facebook)

We built Pharlo to give AI agents a clean way to actually post to social media instead of stopping at "here's a draft."

It's a remote MCP server (Streamable HTTP + OAuth 2.0) and a matching REST API. You connect a channel once via OAuth, then publish now, schedule, update or cancel scheduled posts, retry failed ones, and read analytics — 21 typed MCP tools or the equivalent REST endpoints.

Everything's machine-discoverable: a tools catalog with per-tool credit costs and examples, an OpenAPI 3.0 spec, a .well-known manifest, and an error catalog that maps each failure to a recovery action. Webhooks are HMAC-signed.

Live platforms today are YouTube and Facebook; Instagram and TikTok are planned. Claude is tested; Cursor, Windsurf, ChatGPT, and n8n work over remote OAuth MCP.

Docs: https://pharlo.io/docs · MCP endpoint: https://pharlo.io/_mcp

📣 Pharlo is live — an MCP server + REST API that lets AI agents publish, schedule, and track social posts.
Connect a channel via OAuth, then publish/schedule/retry and read analytics through 21 MCP tools. Live on YouTube & Facebook (Instagram & TikTok planned). Signed webhooks, OpenAPI spec, machine-readable tools catalog.
👉 https://pharlo.io/docs · MCP: https://pharlo.io/_mcp

Pharlo gives AI agents a real publishing surface for social media. It's an MCP server and REST API where an agent connects a channel via OAuth and then publishes, schedules, retries, and reads analytics — across YouTube and Facebook today, Instagram and TikTok planned. It's built to be machine-discoverable (tools catalog, OpenAPI, .well-known manifest, error catalog) with HMAC-signed webhooks and credit-based metering, so agents can publish and verify delivery without a UI.

MCP server, content publishing, social media, scheduling, automation, AI agents, REST API, OAuth, webhooks, analytics, YouTube, Facebook, developer tools