generated: '2026-08-13'
method: searched
source: https://developers.buffer.com/explorer.html
docs:
  - https://developers.buffer.com/explorer.html
  - https://developers.buffer.com/guides/getting-started.html
  - https://developers.buffer.com/guides/cli.html
name: Buffer API Explorer and safe-execution surfaces
summary: >-
  Buffer publishes NO sandbox in the payments sense — there is no test mode, no
  test-vs-live key prefix, no magic test identifiers, no fixture or trigger
  tooling, and no test clock. What it does publish are three real
  safe-execution surfaces: a hosted interactive API Explorer, a client-side
  --dry-run on every CLI mutation, and drafts as a zero-consequence write path.
  Recorded honestly as those three rather than stretched into a sandbox that
  does not exist.
test_mode:
  separate_environment: false
  test_key_prefix: null
  live_key_prefix: null
  note: >-
    One environment, one credential. A personal API key created at
    https://publish.buffer.com/settings/api operates against the real account
    and real connected social channels. Any post created is a real post.
surfaces:
  - name: API Explorer
    kind: hosted console
    url: https://developers.buffer.com/explorer.html
    auth: 'Bring your own API key; the key is shared with the integration setup pages and prefills them.'
    description: >-
      Interactive in-browser query runner. Buffer's Quick Start: "You can run
      this query directly in your browser using the interactive Buffer API
      Explorer, no local setup required."
    consequence: >-
      Executes against the live account. Not a sandbox — reads are safe,
      mutations are real.
  - name: CLI dry run
    kind: client-side validation
    command: 'buffer <group> <command> --json <payload> --dry-run'
    description: >-
      Supported on every mutation. Validates the input locally and prints the
      payload that would be sent, with no API call. Exits 0 only if the input
      parses. Buffer's own agent guidance: "Always dry-run a generated payload
      first; cheap and exits 0 only if input parses."
    counts_against_rate_limit: false
    api_call: false
  - name: Drafts
    kind: reversible write
    mechanism: 'CreatePostInput.saveToDraft: true (CLI: --save-to-draft)'
    description: >-
      Creates a Post with status draft instead of scheduling it. Posting limits
      are not checked and the post will not publish until explicitly scheduled.
      Buffer's own skill text calls this "zero risk" and recommends it for any
      pipeline that might be re-run, since a duplicate draft is cheap to delete
      with posts delete.
    consequence: 'Writes a real record, but never reaches a social network.'
  - name: Schema describe
    kind: local contract inspection
    command: 'buffer schema describe <group> <command> --output json'
    description: >-
      Returns the full JSON Schema for a --json payload, including per-service
      nested metadata, defaultFields and selectableFields. Local only.
    api_call: false
  - name: buffer doctor
    kind: environment check
    command: buffer doctor
    description: >-
      Checks Node version, config, token, default organization, network
      reachability, rate-limit headroom and CLI freshness.
test_values:
  cards: null
  bank_accounts: null
  magic_identifiers: null
  note: >-
    None published. Buffer's pitfalls guidance is the opposite — "Never guess
    channel IDs. Always fetch with buffer channels list. IDs look
    pseudo-random; the API will accept malformed IDs and reject at execute time
    with vague errors." There is nothing to hardcode.
time_simulation:
  test_clock: false
  note: >-
    No time simulation. Scheduling is real: dueAt must be a real ISO-8601
    timestamp with a real offset, and a scheduled post really publishes.
rate_limit_interaction:
  dry_run_free: true
  local_commands_free: [schema, config, doctor, context, '--help', '--version']
  detail: rate-limits/buffer-rate-limits.yml
gaps:
  - >-
    No sandbox account or test organization. A developer building against
    Buffer must connect real social channels, or work exclusively in drafts.
  - >-
    Drafts are the de facto test path, but they still consume idea/post storage
    against the organization's plan limits.
x-evidence:
  - {fetched: '2026-08-13', url: 'https://developers.buffer.com/explorer.html', http_status: 200}
  - {fetched: '2026-08-13', url: 'https://developers.buffer.com/guides/getting-started.html', http_status: 200}
  - {fetched: '2026-08-13', url: 'https://developers.buffer.com/guides/cli.html', http_status: 200}