The posting layer for social media products
Pharlo is the API behind scheduling tools, creator platforms, and AI agents. Connect YouTube and Facebook once — we absorb every API change, quota cut, and OAuth update so your team ships the product, not the plumbing
Three calls. Zero SDKs to maintain
One REST + MCP API in front of every platform. You ship product; we eat the platform-integration tax
Embed our OAuth flow. Users link their Meta, YouTube, or TikTok accounts in three clicks — we own the tokens, scopes, and refresh dance
POST one job — media, metadata, schedule. We absorb every platform's upload, format, and quota surface so your dashboard stays simple.
Signed webhooks the moment a post lands, fails, or returns analytics. Surface it in your UI, feed your AI, trigger an approval — your call
SaaS builders who'd rather ship than maintain SDKs
Stop maintaining what doesn't differentiate you. Every Meta, YouTube, and TikTok change is our problem — not your roadmap's
Launch with more platforms than your competitors and free the team from rewriting OAuth every quarter. Differentiate on UX, not on plumbing
The MCP surface lets your agent post, schedule, and pull analytics in one tool call. Skip the SDK glue — ship the AI feature instead
Publish-on-approve, rollback, multi-stakeholder review. One API powers the whole flow — across every platform your team has to ship to
Add publishing as a native feature without staffing a platform-integrations team. Connect, post, retrieve analytics — that's the whole surface.
Battle-tested by TheSoul Group
Pharlo is the same posting infrastructure that runs TheSoul Group — one of the world's most-viewed digital content companies, posting daily across YouTube, Meta, TikTok and Snapchat for a decade. When a platform changes, we've already shipped the fix in production. You inherit that operational scar tissue on day one
The whole platform surface, behind one contract
REST for your backend, MCP for your agents. OAuth, quotas, upload, analytics — one schema across every platform we cover
One POST creates a job with media, metadata, and an optional schedule. Same payload shape whether it lands on YouTube, Reels, or TikTok
Managed connect flow per platform. Tokens, scopes, refresh, and rotation handled upstream — your code never sees a 401
Views, watch time, subscriber growth, side-by-side channel comparison — normalized across platforms and served straight to your dashboard
HMAC-signed callbacks for backends, MCP tools for agents. The same publishing surface, exposed the way your stack needs it
Idempotent jobs, exponential backoff, per-platform rate-limit awareness. Quota cuts and outages absorbed before they hit your queue
YouTube and Meta live. TikTok, Instagram, Snapchat shipping. Every new platform we add is one we don't add to your roadmap
Same credits
for REST and MCP
Every MCP tool call costs the same as the REST endpoint behind it. 1 credit = $0.001. No AI surcharge, no per-tool tax, no surprise renegotiations when a platform changes its quota
| Operation | Tier | Credits | Cost |
|---|---|---|---|
| Publish to channel Submit a video for delivery to YouTube, Meta, or any connected platform — webhook included | Publish | 25 credits | $0.025 |
| Retry a failed publish Re-queue a failed assignment without re-processing intake | Retry | 5 credits | $0.005 |
| Live platform stats Real-time views, likes, and engagement pulled directly from YouTube / Meta APIs | Platform | 2–3 credits | $0.002–0.003 |
| Update, cancel, or modify Patch a pending assignment, cancel a scheduled one, or update a connection | Write | 2 credits | $0.002 |
| Read — list or fetch detail Any GET on assignments, connections, or webhook delivery logs | Read | 1 credit | $0.001 |
| Auth, OAuth, account & team management Login, token refresh, organizations, invitations, platform discovery, health checks | Free | Free | $0 |
CEILING(size_GB × 25, 1). MCP tool calls use the exact same credit costs as the
equivalent REST endpoint — no surcharge for AI
integrations
Start with $5 — get 5,000 credits
Enough to publish ~200 standard videos, run thousands of reads, and fully exercise both the REST API and MCP server before you commit to anything bigger