Module ai_api 

Enums

AiWebFetchError

struct for typed errors of method ai_web_fetch

AiWebSearchError

struct for typed errors of method ai_web_search

CallAiSessionToolError

struct for typed errors of method call_ai_session_tool

CompleteAiError

struct for typed errors of method complete_ai

GetAiArtifactShareStatusError

struct for typed errors of method get_ai_artifact_share_status

GetAiSessionShareStatusError

struct for typed errors of method get_ai_session_share_status

GetSharedAiArtifactError

struct for typed errors of method get_shared_ai_artifact

GetSharedAiSessionError

struct for typed errors of method get_shared_ai_session

ListAiSessionBackupsError

struct for typed errors of method list_ai_session_backups

ListAiUsageError

struct for typed errors of method list_ai_usage

ListSharedAiArtifactsError

struct for typed errors of method list_shared_ai_artifacts

PullAiSessionBackupsError

struct for typed errors of method pull_ai_session_backups

PushAiSessionBackupsError

struct for typed errors of method push_ai_session_backups

RecordAiUsageError

struct for typed errors of method record_ai_usage

SetAiSessionSharePublicError

struct for typed errors of method set_ai_session_share_public

ShareAiArtifactError

struct for typed errors of method share_ai_artifact

ShareAiSessionError

struct for typed errors of method share_ai_session

UnshareAiArtifactError

struct for typed errors of method unshare_ai_artifact

UnshareAiSessionError

struct for typed errors of method unshare_ai_session

Functions

ai_web_fetch

Has the search API set in the workspace’s AI settings fetch a page and returns its text as markdown. The page is fetched from the search API’s servers, not Windmill’s.

ai_web_search

Runs a query against the search API (Exa, Tavily, Linkup or DeepSeek) set in the workspace’s AI settings, with the key of its resource. The AI chat uses it when the model’s provider has no built-in web search.

call_ai_session_tool

The in-app chat’s entry to the session toolset the MCP server also serves. Refused for job tokens and scoped tokens.

complete_ai

Sends a system and a user message to the configured provider and returns the text of the answer, with no tools and no streaming. For callers that do not speak a provider’s own API through the proxy, such as the CLI and jobs.

get_ai_artifact_share_status

get_ai_session_share_status

get_shared_ai_artifact

get_shared_ai_session

A public share is served without authentication; any other one only to a member of the workspace.

list_ai_session_backups

list_ai_usage

list_shared_ai_artifacts

Every share still inside its retention window, newest first, without its content. Capped at 500.

pull_ai_session_backups

push_ai_session_backups

record_ai_usage

set_ai_session_share_public

share_ai_artifact

Stores a read-only copy that any member of the workspace can open by id. Sharing the same artifact again updates that copy, keeps its id, and restarts its retention window.

share_ai_session

Stores a read-only snapshot of the session that members of the workspace can open by id, or anyone with the link once it is public. Sharing the same session again updates that snapshot and keeps its id.

unshare_ai_artifact

unshare_ai_session