MCP tools & resources

The server exposes read-only knowledge tools grounded in bundled OpenAPI and docs, plus live REST tools when authenticated. Canonical docs are also available as MCP resources.

Auth differs by mode

On the hosted endpoint, knowledge and live tools require OAuth or a legacy API key. On stdio, knowledge tools work without a key; live tools need ADN_API_KEY. See Connect.

Knowledge tools (read-only)

adn_about

Server version, bundled OpenAPI version, docs source, API base

adn_search_docs

Keyword search over bundled docs and OpenAPI summaries

adn_list_operations

List all REST operations (operationId, method, path)

adn_get_operation

Full OpenAPI definition for one operationId

adn_get_guide

Canonical guide by topic: authentication, upload, processing, playback, webhooks, output-types (alias variant-types), security, compatibility, workflows

adn_list_output_types

The seven output types and which are API-creatable (legacy alias: adn_list_variant_types)

Live API tools

Available when the session is authenticated (hosted OAuth/legacy key, or stdio with ADN_API_KEY).

Organization & keys

adn_get_organization
adn_update_organization

Read/update org (id derived from API key when omitted)

read / write

adn_create_client_key

Mint player / uploader / share key (returned once)

write

Creators

adn_list_creators
adn_get_creator

Read creators

read

adn_create_creator
adn_update_creator

Create/update creator

write

adn_delete_creator

Delete creator

destructive

Collections

adn_list_collections
adn_get_collection

Read collections

read

adn_create_collection
adn_update_collection

Create/update collection

write

adn_delete_collection

Delete collection and its tracks

destructive

adn_create_collection_cover

Mint a fresh collection cover upload URL

write

adn_find_resource

Resolve collection/creator/track by name (exact, ambiguous, or missing)

read

Tracks & processing

adn_list_tracks
adn_get_track

Read tracks; prefer adn_get_processing for readiness

read

adn_get_processing
adn_wait_for_processing

Uncached snapshot and bounded wait (playback, output, settled, cover)

read

adn_create_track

Create track + upload URL; client PUTs bytes

write

adn_update_track

Update track metadata / player fields

write

adn_delete_track

Soft-delete one track

destructive

adn_create_track_cover

Mint a fresh track cover upload URL

write

Upload sessions

adn_create_upload_session
adn_get_upload_session

Manage upload sessions (optional nested track)

write / read

adn_create_track_in_upload_session

write

Play sessions

adn_create_play_session
adn_get_play_session

Mint/read play sessions (output indexes via the variants wire field)

write / read

adn_get_play_track

Signed URLs for a track inside a play session

read

adn_get_play_download

Signed download URL for one output index (downloadable sessions)

read

Output presets

adn_list_output_presets
adn_get_output_preset

List/get org output presets (legacy: adn_list_variants / adn_get_variant)

read

adn_create_track_output

Add a track-specific transcode/preview output (legacy: adn_create_track_variant)

write

MCP resources

Clients that attach context directly can use:

audiodn://openapi.json
audiodn://llms-full.txt
audiodn://guide/{topic} for each guide topic

Example agent flow

An agent helping a user host a podcast might:

Call adn_get_guide with topic upload to learn the canonical multi-step flow.
Create a collection with adn_create_collection.
Create an upload session with adn_create_upload_session — optionally pass a nested track to create the first track in the same call.
Otherwise register the file with adn_create_track_in_upload_session, then PUT bytes to track_upload.upload_url.
Wait with adn_wait_for_processing (target track_playback) — call again if the result is pending.
List presets with adn_list_output_presets and pick real output indexes (for example hq, lq).
Mint a play session with adn_create_play_session — returns a play session id for the ADN Player plus signed file URLs on first_track for a custom player.