# Social Neuron — Agent and Developer Guide

> Social Neuron helps creators and teams plan, create, review, distribute, and analyse social content. Customer-specific performance results improve future suggestions. Public documentation describes product outcomes and callable contracts; private algorithms, prompt construction, provider routing, backend topology, and anti-abuse mechanics are not part of the public interface.

## Product capabilities

Social Neuron connects the content lifecycle in one workspace:

1. Research topics and generate brand-aware ideas.
2. Create text, images, video, voice, avatars, carousels, and storyboards.
3. Review content and apply project approval controls.
4. Schedule or publish to connected social accounts.
5. Read analytics and use those results to improve future suggestions.

Publishing is live for YouTube, TikTok, Instagram, and X. LinkedIn and Bluesky are available in early access. Facebook reconnect remains pending. The current product UI is authoritative when a platform provider changes availability.

## Plans and onboarding

The public website offers Free, Starter, Pro, Team, Agency, and Enterprise plans. API execution begins at Pro. A 14-day developer trial can be created without a card and includes scoped access plus trial credits. Generate keys in Settings → Developer; no sales call is required for trial onboarding.

Credits are usage units, not a fixed USD conversion. Generation cost depends on the requested output and current provider pricing. The authenticated product or tool response shows the applicable estimate before generation.

**How do I invite team members?**
Go to Settings → Organization and invite team members by email. Team member access is available on Team (3 members) and Agency (10 members) plans; Pro is single-seat. Each member gets their own login with role-based permissions.

## Developer interfaces

- Developer portal: https://socialneuron.com/developers
- Documentation index: https://socialneuron.com/docs
- REST guide: https://socialneuron.com/docs/rest-api
- CLI guide: https://socialneuron.com/docs/cli-guide
- MCP endpoint: https://mcp.socialneuron.com/mcp
- Public tool discovery: https://mcp.socialneuron.com/v1/tools
- OpenAPI 3.1: https://mcp.socialneuron.com/v1/openapi.json
- MCP server card: https://mcp.socialneuron.com/.well-known/mcp/server-card.json
- npm package: https://www.npmjs.com/package/@socialneuron/mcp-server
- GitHub repository: https://github.com/socialneuron/mcp-server

Discovery is public. Tool execution requires authentication.

## Authentication

Use OAuth 2.0 with PKCE for interactive agents. Discover endpoints from:

- https://mcp.socialneuron.com/.well-known/oauth-authorization-server
- https://mcp.socialneuron.com/.well-known/oauth-protected-resource

For CLI or server automation, create a scoped API key at https://socialneuron.com/settings/developer and send it as `Authorization: Bearer <token>`. Store credentials in an operating-system credential store or secret manager. Never place credentials in prompts, URLs, logs, screenshots, client-side code, or source control.

## Scopes

| Scope | Purpose |
|---|---|
| `mcp:read` | Read analytics, profiles, balances, and lists |
| `mcp:write` | Generate or update content and plans |
| `mcp:distribute` | Schedule and publish content |
| `mcp:analytics` | Refresh and analyse platform data |
| `mcp:comments` | Read and manage supported-platform comments |
| `mcp:autopilot` | Configure managed automation |
| `mcp:full` | Plan-dependent aggregate scope |

Request the narrowest set needed for the user's task. The live tool listing and OpenAPI document declare the required scope for each operation.

## Structured errors

REST errors use JSON with an error type, message, and recovery guidance.

| Status | Agent action |
|---|---|
| `400` | Correct the request; do not retry unchanged |
| `401` | Authenticate again |
| `402` | Check credits or plan access |
| `403` | Request the required scope or use an authorised project |
| `404` | Verify the tool or resource identifier through discovery |
| `429` | Wait for `Retry-After` |
| `5xx` | Retry a safe, idempotent request once, then surface the failure |

## Safe agent operation

- Discover live schemas before execution.
- Do not guess project or resource identifiers.
- Obtain confirmation for consequential operations when required by tool annotations or the user's instruction.
- Use idempotency guidance for retried writes.
- Poll asynchronous jobs only at the cadence described by the live schema.
- Never expose user content or credentials outside the authorised project.

## Live tool discovery

<!-- BEGIN GENERATED: tools (do not hand-edit — run `npm run build:public-feeds` in mcp-server/) -->

_91 public tools are available._

- Live tool names, descriptions, input schemas, scopes, and safety annotations: https://mcp.socialneuron.com/v1/tools
- OpenAPI 3.1 contract: https://mcp.socialneuron.com/v1/openapi.json
- MCP server card: https://mcp.socialneuron.com/.well-known/mcp/server-card.json

Discovery is public. Tool execution requires OAuth 2.0 or a scoped API key.

<!-- END GENERATED: tools -->

## Trust

- About: https://socialneuron.com/about
- Contact: https://socialneuron.com/contact
- Privacy: https://socialneuron.com/privacy
- Terms: https://socialneuron.com/terms
- Security contact: https://socialneuron.com/.well-known/security.txt
- Sitemap: https://socialneuron.com/sitemap.xml

Canonical agent documentation: https://socialneuron.com/llms-full.txt