Skip to content

v0.9.0 — Browser Agent, Onboarding Wizard, Credential Vault

Latest
Latest

Choose a tag to compare

View all tags
@msichalwe msichalwe released this 21 Mar 15:28
· 1 commit to main since this release
v0.9.0
6f2ecca

What's New

Browser Agent 🌐

  • Autonomous Chrome automation — AI navigates websites, fills forms, clicks buttons, extracts data
  • Multi-tab support with per-agent isolation for parallel browsing
  • Smart page analysis: interactive elements with valid CSS selectors
  • Cookie/session persistence across runs
  • 13 new browser tools: browser_agent_run, browser_get_elements, browser_fill_form, browser_extract_table, browser_scroll, and more

Onboarding Wizard ✨

  • 7-step animated setup: Welcome → Use Cases → Theme → Profile → Account → AI Provider → Done
  • General-purpose (not developer-only): coding, writing, research, learning, productivity, creative
  • Mandatory Oni account creation with server-side MongoDB storage
  • Animated transitions, confetti celebration, progress bar

Credential Vault 🔐

  • AES-256-GCM encrypted storage for API keys, passwords, tokens
  • OS Keychain integration via Electron safeStorage
  • 6 AI tools: credential_save, credential_search, credential_use, credential_get, credential_list, credential_delete
  • Masked display in UI — decrypted values never leave main process

Parallel Sub-Agents 🤖

  • Multiple sub-agents now execute concurrently via Promise.all
  • Each browser sub-agent gets its own dedicated Chrome tab (no conflicts)
  • Both streaming paths (OpenAI/Anthropic + ChatGPT/Responses API) support parallel execution

OpenAI Responses API

  • GPT-5.x / GPT-4.1 / o3 / o4 models use Responses API with previous_response_id caching (40-80% token savings)
  • Native built-in tools: web_search_preview and code_interpreter
  • Schema strictness: strict: true + additionalProperties: false for all tools

Model-Aware Context Management

  • 18 model context windows (128K to 1.05M) with automatic compaction thresholds
  • Accurate token estimation (3.8 chars/token, image tokens, tool overhead)
  • Per-model max output tokens

Self-Management Tools

  • get_platform_info — Version, runtime, OS, capabilities, diagnostics
  • update_config — AI can change theme, permission mode, thinking level
  • self_diagnose — Health checks on all subsystems

Rate Limiting

  • Unified rate limit handling across all 4 API paths
  • Countdown timer UI with progress bar and plan type info

Settings Profile Tab

  • Oni account management (login, profile editing, stats, sign out)
  • Connected to backend API with MongoDB

Other Improvements

  • Provider settings grid redesign with "Coming Soon" badges
  • Mode-aware welcome screen with 72 contextual prompts
  • Mode-filtered sidebar recents
  • Chat background differentiation per theme (--bg-chat)
  • Telegram polling exponential backoff
  • Task dependency parsing crash fix

Links

Getting Started

git clone https://github.com/msichalwe/Onicode.git
cd Onicode
npm install
npm run dev
Assets 2
Loading