Overview The docbotConfigSchema is the Zod schema that validates user configuration files. Users typically provide a partial configuration (DocbotUserConfig) which is merged with defaults to create a ResolvedConfig.
Import import { docbotConfigSchema , defineConfig } from "docbot/config" import type { DocbotConfig , DocbotUserConfig } from "docbot/config"
DocbotConfig The inferred TypeScript type from the Zod schema. All fields are optional.
type DocbotConfig = z . infer < typeof docbotConfigSchema >
Configuration fields Agent behavior settings Maximum number of discovery iterations for file exploration. Must be a positive integer. Default: 6
Model overrides for different AI operations. All model IDs must follow the format provider/model-name. Model for long-context analysis operations. Default: "google/gemini-3-pro-preview"
Model for generating embeddings for semantic search. Default: "openai/text-embedding-3-small"
Model for generating high-quality embeddings. Default: "openai/text-embedding-3-large"
Fast model for quick operations. Default: "openai/gpt-5.2"
Lightweight model for simple tasks. Default: "google/gemini-3-flash"
Model for planning and orchestration. Default: "openai/gpt-5.2"
Model for complex planning tasks. Default: "anthropic/claude-opus-4.5"
Model for generating documentation content. Default: "anthropic/claude-sonnet-4.5"
Default paths for CLI inputs Path(s) to the codebase to document. Can be a single path or array of paths.
Path to the documentation output directory.
Project identifier used for collection naming. Must be lowercase alphanumeric with hyphens. Defaults to sanitized package.json name.
Qdrant vector database configuration Qdrant server URL. Must be a valid URL. Default: "http://127.0.0.1:6333"
Custom manifest path for tracking indexed files. Default: ".docbot/manifest.json"
Custom collection names for vector storage Collection name for code embeddings. Default: docbot_{projectSlug}_code
Collection name for documentation embeddings. Default: docbot_{projectSlug}_docs
Server configuration settings Port number for the development server. Must be between 1 and 65535. Default: 3070
DocbotUserConfig Type alias for the input type accepted by the Zod schema. This is what users provide in their docbot.config.ts files.
type DocbotUserConfig = z . input < typeof docbotConfigSchema >
Usage examples Basic configuration import { defineConfig } from "docbot/config" export default defineConfig ({ projectSlug: "my-project" , models: { planning: "openai/gpt-4o" , prose: "anthropic/claude-sonnet-4.5" , } , })
With custom Qdrant settings import { defineConfig } from "docbot/config" export default defineConfig ({ projectSlug: "my-api" , qdrant: { url: "http://localhost:6333" , collections: { code: "custom_code_collection" , docs: "custom_docs_collection" , }, } , })
With path defaults import { defineConfig } from "docbot/config" export default defineConfig ({ paths: { codebase: [ "./src" , "./lib" ], docs: "./documentation" , } , agents: { discoveryBudget: 10 , } , })
Programmatic validation import { docbotConfigSchema } from "docbot/config" const userConfig = { projectSlug: "test" , models: { planning: "openai/gpt-4o" , }, } // Validate config const result = docbotConfigSchema . safeParse ( userConfig ) if ( result . success ) { console . log ( "Valid config:" , result . data ) } else { console . error ( "Invalid config:" , result . error ) }
All model identifiers must follow the format provider/model-name:
Provider: alphanumeric with hyphens (e.g., openai, anthropic, google)
Model name: alphanumeric with hyphens and dots (e.g., gpt-5.2, claude-sonnet-4.5)
// Valid model IDs "openai/gpt-5.2" "anthropic/claude-sonnet-4.5" "google/gemini-3-flash" // Invalid model IDs "gpt-5.2" // Missing provider "openai:gpt-5.2" // Wrong separator
Helper functions defineConfig Provides TypeScript type hints for configuration objects:
function defineConfig ( config : DocbotUserConfig ) : DocbotUserConfig
gateway Re-exported from the ai package for model configuration:
import { defineConfig , gateway } from "docbot/config" export default defineConfig ({ models: { planning: gateway ( "openai/gpt-4o" ), } , })