RAYU
RAYU

Rayu-CLI Changelog

All notable user-facing changes to Rayu-CLI are documented here, newest first.

2.0.13 - 2026-09-21 (Beta)

  • RayuCode on GitHub — you can now connect RayuCode to GitHub and put it to work on your pull requests. Install it as a GitHub App, pick the repositories you want in scope, and RayuCode reviews incoming PRs right where your team already works — leaving inline findings, summaries, and suggested fixes as review comments, no local setup required
  • Run RayuCode as a GitHub bot in your hub repo — designate a "hub" repository and RayuCode becomes a standing bot there. It listens for new PRs and issues, responds to mentions, and can act as the central review and triage agent for the whole organization — one bot, every repo that points at the hub

2.0.12 - 2026-09-20 (Beta)

  • Orchestrator mode is now the default — the CLI now ships with the orchestrator as the built-in coordination layer. Plan, build, review, and run multi-agent workflows end-to-end from a single session, with parallel/sequential/race policies, retry/fallback, and git worktree isolation all wired in by default
  • Reusable messages in VS Code — RayuCode now supports saved, reusable prompt messages. Build a library of your most-used prompts and invoke them instantly from the IDE chat panel, with variable interpolation for project-specific context
  • Collaborator swarm migrated to orchestrator — the standalone /collaborator_swarm command has been retired and its capabilities fully absorbed into the orchestrator. Domain-specialized agents (backend, frontend, mobile, security, deploy) are now orchestrated through the unified /agent workflow with better task routing and shared context
  • Rayu GitHub setup command restored — /install-github-app now creates Rayu-branded @rayu and automated review workflows, defaults to securely provisioning a repository credential from the signed-in Rayu account (with manual API keys still supported), and uses trusted-actor guards plus publisher-configured Rayu GitHub App and Action endpoints

2.0.1 - 2026-09-14 (Beta)

  • 🚀 Introducing RayuCode — our new VS Code extension! Rayu's full AI coding agent engine is now available directly inside VS Code as RayuCode. Same tools, same providers, same skills — now with a native IDE chat panel, diff viewer, and webview UI. One engine, two products: terminal and IDE
  • Greatly improved model support — expanded the provider catalog with more models across Anthropic, OpenAI, Google Gemini, DeepSeek, Kimi, and OpenAI-compatible endpoints. Better model capability detection, smarter fallbacks, and more accurate context window handling
  • Pay-as-you-go billing — introduced a new pay-as-you-go pricing model alongside subscriptions. Top up your balance and pay only for what you use — no monthly commitment required. Perfect for occasional or variable usage
  • More accurate credit charging — completely reworked the credit calculation pipeline. Token counting now aligns precisely with provider billing (input/output/cache-read/cache-write tokens tracked separately), so you're charged exactly for what the model consumed — no more, no less
  • Subscription management overhaul — redesigned the subscription experience from the ground up: clearer plan comparison, seamless upgrades/downgrades, prorated billing, and a unified billing dashboard in rayu-web. View usage, manage your plan, and track invoices in one place
  • Team SSO (Enterprise) — added SAML/OIDC single-sign-on support for teams and organizations. Admins can configure Google Workspace, Okta, or any OIDC IdP so team members sign in with their existing corporate credentials — no separate Rayu accounts needed
  • Fixed streaming interruptions — resolved an issue where long streaming responses could silently drop tokens or terminate early under high-latency network conditions
  • Fixed credit double-counting on retries — credits are now correctly settled (not just reserved) after each request, eliminating the previous over-charge when the gateway retried a failed upstream call
  • Fixed model fallback loops — prevented infinite fallback cycles when multiple models in the chain were unavailable; the CLI now fails fast with a clear error message
  • Fixed subscription plan change not reflecting immediately — plan entitlements are now refreshed in real-time after a subscription change, removing the previous delay where users had to restart the session
  • Fixed memory leak in long-running sessions — resolved a memory accumulation issue in the Ink renderer that caused slowdowns during multi-hour coding sessions
  • Fixed clipboard paste on Windows Terminal — clipboard image paste now works correctly in Windows Terminal and conhost.exe, not just Windows Terminal with the new engine
  • Fixed slash command autocomplete ordering — commands are now ranked by relevance and frequency instead of alphabetically, so /model, /connect, and /help appear first

1.6.0 - 2026-08-30

  • Introduced the External Agent Orchestrator — /agent command and ExternalAgent tool to discover, launch, adopt, assign work to, and stream other agentic CLIs (Codex, Claude Code, OpenCode, and any ACP agent). Supports orchestration policies (parallel/sequential/race/retry/fallback), git worktree isolation, brokered permissions, and crash recovery
  • Overhauled the Telegram bridge for multi-session support — list sessions with /sessions, switch between them with /switch <n>, and route prompts across processes over the new IPC layer
  • Added Telegram bridge health monitoring with typed poll outcomes, exponential backoff, and clear connection status (connected/reconnecting/disconnected)
  • Added Telegram remote uninstall — a typed device lifecycle operation gated by local opt-in, explicit device targeting, single-use confirmation tokens, and a concurrency lock
  • Added a secure IPC (Inter-Process Communication) layer — Unix socket protocol with per-session token auth enabling cross-session routing and multi-session coordination
  • Added the Rayu API Key provider (rayu) — authenticate with a rayu_sk_live_… key, with gateway validation, credential caching, and device identity registration
  • Added a streamlined first-run setup screen — offers Rayu sign-in or API key entry, with the full provider list one keystroke away via /connect
  • Added per-model image capability detection — knows which (provider, model) pairs accept images, replacing the old provider-wide boolean that caused silent image drops or lost turns
  • Added tool input repair — automatically recovers malformed tool-argument JSON from weaker models (markdown fences, raw newlines, trailing commas, truncation)
  • Added Edit/Write tool key coercion — transparently maps camelCase and synonym keys (e.g. filePath → file_path) so tool calls from non-Claude models work correctly
  • Improved the file state cache — tracks full-file reads explicitly to reduce spurious "file modified since read" warnings after formatters or git operations
  • Added device identity — a stable per-machine identifier for remote lifecycle operations, registered and heartbeated against the backend
  • Rewrote the uninstall system — modular orchestrator with scoped manifests, install-method detection, and a detached helper that finishes cleanup after RAYU exits
  • Added the active model name in the input bar and sidebar for better context during sessions
  • Added comprehensive test coverage for external agents, Telegram, API key auth, image capabilities, tool input repair, file edit coercion, and more

1.5.12 - 2026-07-30

  • Improved the model selection experience — better picker UX and more reliable model resolution when switching providers

1.5.11 - 2026-07-30

  • Fixed a bug where Rayu CLI would hang/stuck when running inside the default macOS Terminal app — Enter and several modified key events were not being delivered when the modifiers-napi dependency was absent

1.5.1 - 2026-07-28

  • Migrated the Telegram bridge for a more user-friendly experience — smoother pairing flow, clearer status messages, and more reliable bot re-pairing when the linked bot changes
  • Improved Telegram message handling and bridge reliability for mobile/remote access

1.5.0 - 2026-07-25

  • Migrated the Rayu-hosted provider to support many more models
  • Added real-time model list updates so newly enabled models on the gateway are picked up without requiring a CLI release
  • Improved provider routing and model resolution for Rayu-hosted models

1.4.479 - 2026-07-23

  • Fixed the Ask User Question tool not working in full manage (full-control) permission mode

1.4.478 - 2026-07-21

  • Improved the Ask User Question tool for all LLM providers
  • Fixed the Gateway upstream error
  • Fixed the plan credit calculation bug

1.4.477 - 2026-07-21

  • Fixed Rayu-hosted provider gateway bugs affecting multiple concurrent requests and routing
  • Hardened the internal Rayu provider path — resolved the gate issue that caused intermittent failures under load
  • Improved gateway request handling and route resolution for Rayu-hosted models
  • Tightened gateway logging for clearer diagnostics on hosted-provider failures

1.4.472 - 2026-07-09

  • Fixed tool search (deferred tool loading) for Rayu third-party providers — tool_reference/defer_loading beta was wrongly enabled for anthropic-compatible and openai-compatible providers, causing deferred tools (WebFetch, TaskUpdate, WebSearch, etc.) to be called with guessed parameters and rejected by the client-side validator
  • Added escape hatch: set ENABLE_TOOL_SEARCH=true to force-enable deferred tools on endpoints that forward the beta
  • Added regression test suite for tool search provider detection (toolSearchProvider.test.ts)
  • Cleaned up stale bug-hunter notes

1.4.471 - 2026-07-09

  • Migrated remaining legacy Claude Code internal references and branding to Rayu across the codebase
  • Fixed Kimi Code 2.7 provider compatibility in the streaming pipeline
  • Ollama provider promoted to primary active provider with improved multi-key configuration support
  • Fixed update, uninstall, and install CLI commands for cross-platform reliability

1.4.470 - 2026-07-09

  • All supported providers now route through Rayu-hosted gateway with full multi-provider compatibility
  • Improved handling of disabled providers — long-running stale provider connections are now properly cleaned up
  • UI polish: refined several terminal components for better readability and consistency
  • Added native context window support for Rayu-hosted models

1.4.466 - 2026-07-08

  • Rebranded from Claude theme to Rayu branding theme with new Goose mascot identity
  • Added /banner and /mascot slash commands for displaying Rayu branding in the terminal
  • Introduced LogoV2 components with animated goose mascot and configurable banner system
  • Migrated legacy Claude-specific modules (feedback, grove, analytics, OAuth) to un-use-code/ archive
  • Removed Telegram payment notification listener and Claude subscription integrations
  • Added mascot image encoding, Unicode rendering, and caching utilities for terminal display
  • Implemented clean posture verification scripts and origin manifest for independent branding audit
  • Added anthropicCompatibleClient for provider-agnostic Anthropic-compatible API consumption

1.3.462 - 2026-07-06

  • Removed DeepSeek Web reverse-scraping provider
  • Added cache read/write credit multipliers for usage tracking
  • Improved gateway graceful shutdown and added related tests
  • Updated Kiro provider to Sonnet 5
  • Improved ABA/QrCode payment flow and Telegram listener reliability

1.3.445 - 2026-06-29

  • Fixed bugs in gateway routing logic for rayu-hosted provider
  • Added tests for gateway routing edge cases

1.3.444 - 2026-06-29

  • Implemented prompt input color theming via theme.ts
  • Improved thinking message display and highlighted thinking text rendering
  • Added structured API error handling in errors.ts
  • Improved context analysis and consistency checks
  • Fixed renderer ghost-character edge cases in log-update.ts

1.3.443 - 2026-06-26

  • Production stability improvements: entitlements, auth session, and gateway routing
  • Fixed preload script and macro values for production launch
  • Improved dashboard page and first-time launch flow in rayu-web

1.3.442 - 2026-06-24

  • Improved interrupt message handling during streaming queries
  • Added memory pressure guard to prevent OOM crashes
  • Added heap limit re-exec mechanism for large sessions
  • Improved curated provider model list and provider configuration

1.3.441 - 2026-06-24

  • Refactored backend payment integration to support Bakong (KHQR) payments
  • Added Telegram listener for ABA payment notifications
  • Updated payment DTOs and controller for new payment flow

1.3.440 - 2026-06-22

  • Fixed missing TypeScript type modules across SDK and component types
  • Added FPS tracker, memory probe, and progress coalescing utilities
  • Added interactive heap dump monitor
  • Fixed copilot auth, message handling, and OpenAI adapter test coverage

1.3.439 - 2026-06-21

  • Implemented basic subscription plan catalog and entitlements enforcement
  • Added paid feature gating for image generation and video generation tools
  • Added usage limits for free plan users (rayuFeatureUsage.ts, paidFeatureGate.ts)
  • Refactored thinking message display in assistant responses
  • Improved gateway credit limiter, entitlements cache, and admin panel in rayu-web

1.3.438 - 2026-06-17

  • Added Hugging Face (huggingface) provider preset using Hugging Face Inference Providers endpoint (https://router.huggingface.co/v1) with HF_TOKEN
  • Implemented feature rate limits, entitlements, and Copilot client authentication
  • Added/updated integrations for image-editor, video-generation, and image-generation tools
  • Implemented and improved Admin Dashboard features (analytics, payments, plans, users) in rayu-web and rayu-backend

1.3.437 - 2026-06-17

  • Fixed Kiro provider thinking payload construction and context base

1.3.436 - 2026-06-17

  • Reworked and renamed the Planning & Research subagent from "PA-AGENT" to "planner"
  • Optimized parallel exploration of codebase in planner subagent (concurrent dispatch of Explore subagents)
  • Improved collaborator swarm and persistent agent memory syncing/loading
  • Added UI spinner and status display improvements for agent tasks

1.3.435 - 2026-06-17

  • Published new stable CLI version with package improvements

1.3.434 - 2026-06-16

  • Fix the bug with the kiro provider context
  • Rebrand from Claude icon to Rayu draft icon
  • Add /brandmark slash command to customize brand mark glyph and loading-spinner style

1.3.433 - 2026-06-15

  • Implement the doubleword provider to fetch all model

1.3.432 - 2026-06-15

  • fixed Bug of thinking

1.3.431 - 2026-06-15

  • fix kiro context base on the original context 1M

1.3.430 - 2026-06-15

  • The final update the rayu cli free of use.
  • Added new provider with kiro provider login through api key and the kiro OAuth
  • Add Rayu OAuth for preview first.

1.3.429 - 2026-06-14

  • Added FIX thinking to all provider
  • we improve the thinking to for provider and make the speed bettern 10%

1.3.428 - 2026-06-13

  • Added the thinking status display to all agent subagents and the collaborator agent

1.3.427 - 2026-06-13

  • Completely updated the documentation to properly reflect Rayu as an independent CLI, removing legacy Claude Code specific references
  • Improved the thinking UI and status displays
  • Fixed a bug with Kimi provider thinking output
  • Fixed a bug causing blocked thinking through the Bedrock provider

1.2.25 - 2026-06-11

  • Extended thinking and effort now work on any provider, not just Claude — type ultrathink to trigger deep reasoning on NVIDIA, Gemini, DeepSeek, and other OpenAI-compatible models
  • /effort max is no longer labeled "Opus only" and applies to whatever model you have connected
  • Added local /ultraplan: deep multi-agent planning that explores your codebase in parallel and produces a step-by-step plan for approval — runs entirely on your own provider
  • Added local /ultrareview: a deep bug-hunt that gathers your branch diff and dispatches parallel review subagents to find and verify real bugs — runs entirely on your own provider
  • Enabled the "ultrathink" keyword highlight and per-turn token budgets (type +500k or use 2M tokens)
  • Sub-agents can now use the built-in Explore and Plan helpers for faster, deeper investigation
  • "What's new" now shows Rayu's own changelog, and an "Update available" notice appears when a newer version is published to npm

1.2.24 - 2026-06-10

  • Fixed a spurious "Interrupted by user" error that appeared when a parallel tool timed out or a sibling command failed — interruptions are now only reported when you actually cancel
  • /ide now detects VS Code when connected through the integrated terminal
  • Clipboard image paste now works on Wayland (Linux)
  • Reworked /review_detail into an interactive diff viewer
  • /undo all reverts every pending file change at once
  • Renamed /stickers to /contact_me