Skip to content

v0.1.1 - Smart Model Combos, Failover Cascades & Robust Resilience

Choose a tag to compare

View all tags
@seaavey seaavey released this 18 Aug 09:56
· 230 commits to main since this release
v0.1.1
a2e7edd
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

⚡ SRouter v0.1.1 — Smart Model Combos, Token Saver & Resilient Gateway

SRouter v0.1.1 is a major feature and stability release introducing Smart Model Fallback Cascades & Priority Combo Routing, an intelligent Multi-Stage Token Saver Engine, full Anthropic Messages API protocol compliance, and resilient error recovery across all gateway routes.

🌟 What's New in v0.1.1

🔀 Smart Model Combos & Automated Failover Cascades (/combo)

  • Grouped Visual Pipeline: Intuitive multi-tier cascade visualization (Step 1 Primary ➔ Step 2 Failover ➔ Step 3 Backup) with provider branding logos and model capability badges (Vision 👁️ & Reasoning 🧠).
  • macOS-Style Dialog & Model Picker: Streamlined modal dialog with live regex validation, model priority reordering, and model selector modal categorized by provider.
  • 1-Click Starter Templates: Quick presets for Flagship Fallback Chain, High Throughput & Speed, and Deep Reasoning Mix.
  • Zero Idle Latency: Combo virtual aliases directly route to target model chains without candidate 0 execution delays.
  • Case-Insensitive Resolution: Fallback matching supports case-insensitive model identifiers (claude matches Claude).
  • Resilient Fallback Triggers: Automated failover on HTTP 429, 403, 5xx, missing provider drivers, rate limit exhaustion, and upstream connection failures.
  • Quick Devtools: Direct "Test in Playground" (/playground?model=...) and instant "Copy cURL" command export.

✂️ Multi-Stage Token Saver Engine (/settings/token-saver)

  • Prompt Token Compaction: Rule-based token optimization reducing prompt overhead and upstream billing cost before LLM execution.
  • Customizable Rules: Whitespace compaction, redundant code comments pruning, and repetitive string compression.
  • Dashboard Telemetry: Dedicated UI showing estimated token savings, reduction percentages, and audit metrics.

🔌 New Upstream AI Providers & Protocol Bridges

  • Expanded Provider Catalog: Native executors for GoRouter, BluesMinds, SeekAI, TabiToken, Neosantara, and Qoder (Alibaba Cloud PKCE OAuth flow with WAF bypass).
  • Anthropic Messages API: Native POST /v1/messages & POST /messages support with bidirectional tool calling and SSE streaming (compatible with Claude Code CLI, Cursor, and official Anthropic SDKs).

🛡️ Robust Validation & Error Handling

  • Graceful JSON Parsing: Handles malformed JSON payloads and empty bodies returning standard OpenAI invalid_request_error (HTTP 400) without throwing unhandled Hono HTTPException.
  • Global Error Normalization: Centralized error interceptors across primary API listeners and port 1455 OAuth servers.

🔗 Pull Requests & Related Changes

PR / Issue Description Author
#23 feat(routing): introduce Smart Model Fallback Cascades & Priority Combo Routing @seaavey
#22 feat(token-saver): add multi-stage token saver engine and web dashboard @seaavey
#21 feat(gateway): Multi-Provider Priority Failover & Fallback Rules Engine @seaavey
#20 fix(web): unbreak dashboard onConnect & handle multi-tenant OAuth refresh @Sienz16
#19 fix(models): speed up provider model discovery with concurrent background sweeps @Sienz16
#18 fix(docker): make pnpm workspace caching resilient @Sienz16
#17 feat(auth): add first-run admin onboarding & JWT session management @Sienz16
#16 fix(docker): copy per-package manifests for faster Turbo caching @Sienz16
#14 feat(api): add Anthropic Messages API support with SSE delta streaming @seaavey
#13 feat(provider): add Qoder OAuth PKCE & Alibaba Cloud upstream proxy @seaavey
#10 feat: add Neosantara provider integration @seaavey

📌 Closed Issues

💖 Special Thanks & Contributors

Huge thanks to everyone who contributed to making this release possible!

  • @seaavey (Lead Architect & Core Maintainer) — Gateway architecture, fallback cascades engine, model combos UI, Token Saver engine, and provider integrations.
  • @Sienz16 — Admin onboarding auth flow, provider discovery concurrency optimizations, and Docker workspace caching enhancements.

Full Changelog: v0.1.0-beta...v0.1.1

Contributors

seaavey and Sienz16
Assets 2
Loading