Configuration Guide

DevDiff supports flexible configuration using a .devdiff.config.js file (recommended), .devdiff.config.json, or other .devdiffrc configurations in the root directory of your repository.

---

Complete Configuration Template

Here is a full schema reference detailing all available options in .devdiff.config.js format:

// .devdiff.config.js
export default {
  version: "1.0.6",

  // ── AI CONFIGURATION ──
  ai: {
    providers: [
      {
        name: "ollama-local",
        url: "ollama://qwen2.5-coder:7b",
        priority: 1,
        timeout: 60000, // 60 seconds
        maxTokens: 4096,
      },
      {
        name: "openai-cloud",
        url: "openai://gpt-4o-mini",
        apiKey: process.env.OPENAI_API_KEY,
        priority: 2,
        timeout: 30000,
      },
    ],
    routing: {
      strategy: "priority", // priority, cost-aware, latency, capability
      complexityThreshold: 0.5,
      localOnly: false,
      maxDailyCost: 0.5, // USD limit
    },
  },

  // ── SECURITY CONTROLS ──
  security: {
    disableShellAccess: false,
    disableNetworkAccess: false,
    allowedShellCommands: ["git", "ollama", "which"],
    auditEncryption: true,
    fipsMode: false,
  },

  // ── PRIVACY CONTROLS ──
  privacy: {
    auditLogRetention: 30, // retain logs for 30 days
    autoDeleteAuditLogs: false,
    dataMinimization: "strict", // strict, moderate, relaxed
    blockExternal: ["**/.env", "**/secrets/**", "**/keys/**"],
  },

  // ── PERFORMANCE TUNING ──
  performance: {
    maxMemoryMB: 256, // limit peak heap memory to 256MB
    maxFilesPerAnalysis: 500,
    useIncrementalDiffing: true,
    cacheASTRuns: true,
  },

  // ── CONTEXT WINDOW LIMITS ──
  context: {
    maxContextTokens: 4000,
    overflowStrategy: "summarize", // truncate, summarize, split, mcp
  },

  // ── AGENTIC WORKSPACE SETTINGS ──
  agentic: {
    enabled: true,
    autoStart: true,
    autoAnalyzeOnCommit: false,
    notificationMode: "minimal", // silent, minimal, verbose
    maxAutoAnalysesPerHour: 20,
  },

  exclude: [
    "**/node_modules/**",
    "pnpm-lock.yaml",
    "package-lock.json",
    "dist/**",
    "*.log",
  ],

  cache: {
    enabled: true,
    path: ".devdiff/cache.json",
  },

  format: "markdown",
};

---

Detailed Option Breakdown

AI & Routing Parameters

Option	Type	Default	Description
ai.providers	Array	Required	List of model provider endpoints and credentials. Supports local (ollama://) and cloud (openai://, gemini://, anthropic://).
ai.routing.strategy	string	"priority"	Strategy used by the Intelligent Router. Options: 'priority', 'cost-aware', 'latency', 'capability'.
ai.routing.complexityThreshold	number	0.5	Complexity value above which changes trigger escalation to higher capacity models (0-1).
ai.routing.localOnly	boolean	false	When true, prevents the router from selecting any external cloud providers regardless of complexity.
ai.routing.maxDailyCost	number	0.50	Maximum daily spend on cloud AI services (in USD) to prevent surprise billing.

---

Security Parameters

Option	Type	Default	Description
security.disableShellAccess	boolean	false	Disables execution of all shell commands, running purely in static analysis mode.
security.disableNetworkAccess	boolean	false	Block all outbound connections, forcing local-only Ollama runs.
security.allowedShellCommands	Array<string>	["git", "ollama", "which"]	Whitelist of executable commands.
security.auditEncryption	boolean	true	Encrypt audit logs at rest using AES-256-GCM.
security.fipsMode	boolean	false	Run with FIPS 140-2 validated cryptography libraries.

---

Privacy & Data Minimization

Option	Type	Default	Description
privacy.auditLogRetention	number	30	Number of days to retain system audit logs.
privacy.autoDeleteAuditLogs	boolean	false	Auto-prune logs older than auditLogRetention days.
privacy.dataMinimization	string	"strict"	How aggressively to minimize payload data sent to the AI. Options: 'strict', 'moderate', 'relaxed'.
privacy.blockExternal	Array<string>	["**/.env", "**/secrets/**", "**/keys/**"]	Glob patterns of files that must NEVER be sent to external cloud APIs.

---

Performance & Memory Tuning

Option	Type	Default	Description
performance.maxMemoryMB	number	256	Heap memory cap (in MB) for DevDiff processes.
performance.maxFilesPerAnalysis	number	500	Max count of files allowed in a single generate run.
performance.useIncrementalDiffing	boolean	true	Enables indexing to check changed blocks incrementally.
performance.cacheASTRuns	boolean	true	Cache AST parses to optimize CPU usage.

---

Context Window & Agentic Mode

Option	Type	Default	Description
context.maxContextTokens	number	4000	Token limit for prompts.
context.overflowStrategy	string	"summarize"	How to handle tokens exceeding limit. Options: 'truncate', 'summarize', 'split', 'mcp'.
agentic.enabled	boolean	true	Enable or disable agentic IDE workspace integration.
agentic.autoStart	boolean	true	Start agentic daemon automatically when editor opens.
agentic.notificationMode	string	"minimal"	Alert verbosity. Options: 'silent', 'minimal', 'verbose'.

---

Editor Autocomplete (IntelliSense)

DevDiff CLI can generate JSON Schema files to enable auto-complete, parameter descriptions, and validation directly in your editor:

# Generate the VS Code settings schema
devdiff schema

This creates a local .vscode/devdiff-schema.json file. Follow the CLI directions to add the schema mapping in your .vscode/settings.json.