readModelsCache returns the cached model list for (providerName, fullURL). When maxAge > 0 the entry must be younger than maxAge; maxAge <= 0 accepts

Claude message conversion

APIKeyMiddleware validates API key for protected endpoints. Swagger documentation endpoints (/swagger/*) are exempt from authentication to allow users

AzureDeploymentMiddleware transforms Azure OpenAI API paths to include the deployment name. Azure requires URLs like: openai/deployments/{deployment-n

BuildEndpoint constructs the Azure OpenAI API endpoint URL API version reference: https://learn.microsoft.com/azure/ai-services/openai/reference

Execute runs an extension with the given operation and value string name: the registered name of the extension operation: the operation to perform val

GetGeminiVoices returns the current list of supported Gemini TTS voices This list is maintained based on official Google Gemini documentation https://

GetLatestTag returns the name of the most recent tag by committer date

GrabMetadataForURL retrieves metadata for a Spotify URL (show or episode).

ListModels returns the list of models available for this backend

NewClient creates a new LM Studio client with default configuration.

NewContextsHandler creates a new ContextsHandler

ParseResponse parses Bedrock API response (Anthropic content blocks)

ScanFiles scans specific files and returns a JSON representation

addImageGenerationTool adds the image generation tool to the request if needed

buildModelNameFull adds the "models/" prefix for API calls

doRequest performs an authenticated request to the Spotify API.

extractAndValidateVideoId extracts a video ID from the given URL and validates that the URL points to a video rather than a playlist-only resource. It

filterConversationalModels returns only models suitable for text generation/chat

retrieves a pattern from the database by name

getMaxTokens returns the max output tokens to use for a request

Save pattern outputs and starred outputs to files. Error handling: - Creates output directory if it doesn't exist - Handles file write pe

sendNotification sends a desktop notification about command completion. When truncating the result for notification display, this function counts Uni

setHSTS sets the Strict-Transport-Security header. Each validation 400 sends it, the same as the chat BindJSON 400 path.

supportsImageGeneration checks if the given model supports the image_generation tool

writeModelsCache persists a successfully fetched model list. Empty lists are not cached so a transient empty response does not stick for the whole TTL

BuildEndpoint appends the /openai/ suffix to the given base URL.

ExtractDeploymentFromBody reads the model field from the JSON request body and restores the body for subsequent use.

FetchFilesFromRepo clones a git repo and extracts files from a specific folder. It tries go-git first, and falls back to the git CLI if available.

FetchPlaylistVideos fetches all videos from a YouTube playlist.

GetAbsolutePath resolves a given path to its absolute form, handling ~, ./, ../, UNC paths, and symlinks.

GetEpisodeMetadata retrieves metadata for a Spotify episode.

GetNames overrides StorageEntity.GetNames to include custom patterns directory

GetShowMetadata retrieves metadata for a Spotify show (podcast).

GrabTranscript retrieves the transcript for the specified video ID using yt-dlp. The language parameter specifies the preferred subtitle language code

Init Initialize flags. returns a Flags struct and an error

IsConfigured returns true if a custom patterns directory has been set

NewExtensionExecutor creates a new executor instance It requires a registry to verify extensions

ParseDeployments splits a comma-separated deployment string into a slice, trimming whitespace and discarding empty entries.

ProcessExtension handles template processing for extension directives

RemoveExtension handles the rmextension flag action

SaveCommitPRMappings saves SHA→PR mappings for all commits in PRs

Send sends a notification using the configured provider

SetEnableXSearch toggles whether an additional xAI "x_search" tool entry is appended when Search is enabled. Non-xAI providers should leave this false

SetWebSearchToolName overrides the default "web_search_preview" tool name emitted on the Responses API when Search is enabled. Pass an empty string to

addAuthHeader adds the authorization header to a request.

buildGeminiConfig creates the generation config for Gemini models following the gemini.go pattern for feature parity

buildPRWithCommits fetches commits and constructs a PR object from a GitHub API response

buildVendorsMap groups the model names by vendor name for the response. A vendor with no models gets an empty slice, not a nil slice, because a nil sl

ensureIncomingDir creates the incoming directory if it doesn't exist

getPreferredLocale returns the best locale to use based on user preferences. Priority order: 1. Explicit language flag (if provided) 2. System environ

getStrategyDir returns the path to the strategies directory