buffer-api

name	buffer-api
version	1.1.0
description	Schedule, create, and manage social media posts via the Buffer GraphQL API. Use this skill when the user wants to post to social media through Buffer, schedule posts, create ideas, list channels, retrieve posts, or manage their Buffer queue. Triggers on: Buffer, schedule a post, publish to social media, Buffer API, social media queue, Buffer channels, create Buffer post, Buffer idea, or any mention of posting to LinkedIn/Twitter/Instagram/Facebook/Bluesky/Threads/Pinterest/Mastodon/YouTube/TikTok via Buffer.

Buffer GraphQL API

Schedule, create, and manage social media posts through Buffer's GraphQL API.

Prerequisites

• A Buffer account (any plan) with verified email
• An API key from Buffer API Settings
• Store the key as BUFFER_API_KEY environment variable

Paid accounts can generate up to 5 API keys. Free accounts get 1. TikTok is not supported via the API at this time.

Endpoint

POST https://api.buffer.com
Content-Type: application/json
Authorization: Bearer $BUFFER_API_KEY

All requests are GraphQL POST requests to this single endpoint.

Authentication

Every request needs the Authorization: Bearer header:

curl -X POST 'https://api.buffer.com' \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $BUFFER_API_KEY" \
  -d '{"query": "{ account { organizations { id } } }"}'

Supported Operations

Operation	Type	Description
account	Query	Get authenticated user info
channel	Query	Fetch single channel by ID
channels	Query	Fetch all channels for an org
post	Query	Fetch single post by ID
posts	Query	Fetch posts with filters, sorting, pagination
dailyPostingLimits	Query	Check posting limits per channel per date
createPost	Mutation	Create and schedule a post
deletePost	Mutation	Delete a post by ID
createIdea	Mutation	Save an idea to the ideas board

Setup Flow

Before creating posts, you need your organization ID and channel IDs. Run these queries in order:

1. Get Organizations

query GetOrganizations {
  account {
    organizations {
      id
    }
  }
}

2. Get Channels

query GetChannels {
  channels(input: { organizationId: "YOUR_ORG_ID" }) {
    id
    name
    service
    type
  }
}

Filter by service type:

query GetFilteredChannels {
  channels(
    input: {
      organizationId: "YOUR_ORG_ID"
      filter: { services: [linkedin] }
    }
  ) {
    id
    name
    service
    type
  }
}

Service enum values: facebook, twitter, instagram, linkedin, pinterest, googlebusiness, mastodon, tiktok, bluesky, youtube, threads, startpage

3. Get a Single Channel

query GetChannel {
  channel(input: { channelId: "YOUR_CHANNEL_ID" }) {
    id
    name
    service
    type
    metadata {
      ... on LinkedInMetadata { handle }
      ... on TwitterMetadata { handle }
      ... on FacebookMetadata { facebookPageName }
      ... on InstagramMetadata { handle }
    }
  }
}

Creating Posts

Text Post

mutation CreatePost {
  createPost(input: {
    text: "Hello from the Buffer API!",
    channelId: "YOUR_CHANNEL_ID",
    schedulingType: automatic,
    mode: addToQueue
  }) {
    ... on PostActionSuccess {
      post {
        id
        text
        status
      }
    }
    ... on MutationError {
      message
    }
  }
}

Image Post

Same as text post, with assets.images:

mutation CreateImagePost {
  createPost(input: {
    text: "Check out this image!",
    channelId: "YOUR_CHANNEL_ID",
    schedulingType: automatic,
    mode: addToQueue,
    assets: {
      images: [
        { url: "https://example.com/image.jpg" }
      ]
    }
  }) {
    ... on PostActionSuccess {
      post {
        id
        text
        assets { id mimeType }
      }
    }
    ... on MutationError {
      message
    }
  }
}

Scheduled Post (Specific Time)

Use mode: customSchedule with a dueAt ISO 8601 timestamp:

mutation CreateScheduledPost {
  createPost(input: {
    text: "Posting at a specific time",
    channelId: "YOUR_CHANNEL_ID",
    schedulingType: automatic,
    mode: customSchedule,
    dueAt: "2026-04-01T14:00:00.000Z"
  }) {
    ... on PostActionSuccess {
      post { id text status }
    }
    ... on MutationError { message }
  }
}

Share Now

Use mode: shareNow:

mutation ShareNow {
  createPost(input: {
    text: "Publishing immediately!",
    channelId: "YOUR_CHANNEL_ID",
    schedulingType: automatic,
    mode: shareNow
  }) {
    ... on PostActionSuccess {
      post { id text status }
    }
    ... on MutationError { message }
  }
}

Share Next

Use mode: shareNext to push the post to the front of the queue:

mutation ShareNext {
  createPost(input: {
    text: "This goes next in the queue",
    channelId: "YOUR_CHANNEL_ID",
    schedulingType: automatic,
    mode: shareNext
  }) {
    ... on PostActionSuccess {
      post { id text status }
    }
    ... on MutationError { message }
  }
}

CreatePostInput Fields

Field	Type	Required	Description
text	String	Yes	Post content
channelId	ChannelId	Yes	Target channel
schedulingType	Enum	Yes	automatic or notification
mode	Enum	Yes	addToQueue, shareNow, shareNext, customSchedule
dueAt	DateTime	For customSchedule	ISO 8601 timestamp
assets	AssetsInput	No	Images, videos, documents, links

AssetsInput

assets: {
  images: [{ url: "https://..." }]
  videos: [{ url: "https://...", title: "My Video" }]
  documents: [{ url: "https://..." }]
  link: { url: "https://..." }
}

Creating Ideas

Save content ideas to the Buffer ideas board:

mutation CreateIdea {
  createIdea(input: {
    organizationId: "YOUR_ORG_ID",
    content: {
      title: "Blog post idea"
      text: "Write about the new GraphQL API features"
    }
  }) {
    ... on Idea {
      id
      content { title text }
    }
  }
}

Retrieving Posts

Get Scheduled Posts

query GetScheduledPosts {
  posts(
    input: {
      organizationId: "YOUR_ORG_ID",
      filter: { status: [scheduled] },
      sort: [{ field: dueAt, direction: asc }]
    }
  ) {
    totalCount
    edges {
      node {
        id
        text
        status
        createdAt
        dueAt
        channelId
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

Get Sent Posts for a Channel

query GetSentPosts {
  posts(
    input: {
      organizationId: "YOUR_ORG_ID",
      filter: {
        status: [sent],
        channelIds: ["YOUR_CHANNEL_ID"]
      }
    }
  ) {
    edges {
      node {
        id
        text
        createdAt
        channelId
      }
    }
  }
}

Get Posts with Assets

query GetPostsWithAssets {
  posts(
    input: {
      organizationId: "YOUR_ORG_ID",
      filter: { status: [sent], channelIds: ["YOUR_CHANNEL_ID"] }
    }
  ) {
    edges {
      node {
        id
        text
        assets {
          thumbnail
          mimeType
          source
          ... on ImageAsset {
            image { altText width height }
          }
        }
      }
    }
  }
}

Paginated Posts

Use first and after for cursor-based pagination:

query GetPaginatedPosts {
  posts(
    input: { organizationId: "YOUR_ORG_ID" },
    first: 10,
    after: "CURSOR_FROM_PREVIOUS_PAGE"
  ) {
    edges {
      node { id text status }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
    totalCount
  }
}

PostStatus Enum Values

draft, scheduled, sent, error, deleted, approval_pending, approval_rejected

PostSortableKey Enum Values

dueAt, createdAt

Deleting Posts

mutation DeletePost {
  deletePost(input: { postId: "POST_ID" }) {
    ... on DeletePostSuccess {
      post { id status }
    }
    ... on MutationError { message }
  }
}

curl Recipes

Quick post to queue

curl -X POST 'https://api.buffer.com' \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $BUFFER_API_KEY" \
  -d "$(jq -n --arg text "Hello from the API" --arg ch "$BUFFER_CHANNEL_ID" '{
    query: "mutation CreatePost($input: CreatePostInput!) { createPost(input: $input) { ... on PostActionSuccess { post { id text } } ... on MutationError { message } } }",
    variables: { input: { text: $text, channelId: $ch, schedulingType: "automatic", mode: "addToQueue" } }
  }')"

List scheduled posts

curl -X POST 'https://api.buffer.com' \
  -H 'Content-Type: application/json' \
  -H "Authorization: Bearer $BUFFER_API_KEY" \
  -d "$(jq -n --arg org "$BUFFER_ORG_ID" '{
    query: "query($input: PostsInput!) { posts(input: $input, first: 20) { edges { node { id text dueAt channelId } } } }",
    variables: { input: { organizationId: $org, filter: { status: ["scheduled"] }, sort: [{ field: "dueAt", direction: "asc" }] } }
  }')"

Agentic Workflow & Vibe Coding

• Iterative Scheduling: Do not expect complex GraphQL payloads to perfectly schedule multi-asset posts on the first try. Draft the payload, test it as a draft or send it to the ideas board, isolate any specific schema or media attachment errors, refine ONE field at a time, and re-test until the post format is correct.
• Vibe Coding: Commit your working GraphQL queries or scheduling scripts locally before attempting destructive operations or bulk queue modifications.

Error Handling

Mutation Errors (Recoverable)

Returned as typed union members. Always include ... on MutationError { message } in mutations:

... on PostActionSuccess { post { id } }
... on MutationError { message }

Specific error types: QueueLimitError, PostAlreadyExistsError, ValidationError, LimitReachedError, InvalidInputError, UnauthorizedError, VoidMutationError

Query Errors (Non-recoverable)

Returned in the standard GraphQL errors array with extension codes:

Code	Meaning
NOT_FOUND	Resource does not exist
FORBIDDEN	No permission
UNAUTHORIZED	Auth required or invalid
UNEXPECTED	Server error
RATE_LIMIT_EXCEEDED	Too many requests

Rate Limits

Client Type	Limit
Third-party clients	100 requests / 15 min
Unknown/unauthenticated	50 requests / 15 min
Account overall (all clients)	2,000 requests / 15 min

Response headers: RateLimit-Limit, RateLimit-Remaining, RateLimit-Reset

On 429, use the retryAfter value (seconds) from the error response.

Query Limits

Limit	Value
Max query complexity	175,000 points
Max query depth	25 levels
Max aliases	30
Max directives	50
Max tokens	15,000

Supported Channels

Facebook, Twitter/X, Instagram, LinkedIn, Pinterest, Google Business, Mastodon, TikTok (not via API), Bluesky, YouTube, Threads, Start Page

Agentic OS Integration

If the current project root contains an .agent/ directory, this skill MUST participate in the Agentic OS shared-memory model.

At the end of your execution, check for .agent/state/last-run.json. If it exists, append or update the file using its required schema to log your run. Ensure you capture your runtime (agent_runtime), skill_executed, a concise summary, decisions, and next_steps.

Reference

• Developer Docs
• API Reference
• GraphQL Explorer
• API Changelog
• API Status
• Discord