v0.9.7 — brcc logs + fallback payment fix

Features

• brcc logs command: View debug logs without digging through files
  • brcc logs — show last 50 entries
  • brcc logs -f — follow in real time (like tail -f)
  • brcc logs -n 100 — show last 100 entries
  • brcc logs --clear — delete log file
  • Auto-rotates at 10MB so it won't eat up disk space
• Always-on logging: [brcc] messages always written to ~/.blockrun/brcc-debug.log — no need for --debug flag for basic logs

Bug Fixes

• Fallback + payment mismatch: When fallback switches to a different model and the backend returns 402 (payment required), the payment handler now uses the correct fallback model body instead of the original failed model body. This was causing fallback to silently fail on paid models.

Upgrade

sudo npm update -g @blockrun/cc

v0.9.6 — Fix login prompt (ANTHROPIC_AUTH_TOKEN)

Bug Fixes

• Login prompt fix: Use ANTHROPIC_AUTH_TOKEN instead of ANTHROPIC_API_KEY to prevent Claude Code from showing login prompt when launched via brcc start (thanks @0xCheetah1, #2)
• Consistent env var in all output messages (proxy-only mode, error fallback)

Upgrade

sudo npm update -g @blockrun/cc

v0.9.5 — Fix fallback 400 errors

Bug Fixes

• Fallback 400 errors: Removed virtual routing profiles (blockrun/auto, blockrun/eco) from fallback chain — backend doesn't recognize these, causing 400 loops. Fallback now uses concrete models: deepseek/deepseek-chat → google/gemini-2.5-flash → nvidia/nemotron-ultra-253b
• Safety filter: buildFallbackChain() now strips routing profiles to prevent them from ever reaching the backend
• brcc start automation: Smarter claude binary detection — searches PATH + common install locations. Falls back to printing manual env vars instead of crashing

Docs

• Use blockrun.ai/brcc-install short URL in README
• Added CHANGELOG

Upgrade

sudo npm update -g @blockrun/cc

Full changelog: https://github.com/BlockRunAI/brcc/blob/main/CHANGELOG.md

v0.9.4 — Fix model routing for default startup

Bug Fix

Fixed: brcc start (without --model) was passing Claude Code's native model IDs (e.g. claude-sonnet-4-6-20250514) directly to the BlockRun backend, causing "Unknown model" errors.

Root cause: The proxy only overrode the model when --model was explicitly passed on the CLI. But Claude Code always sends its own native Anthropic model IDs, which don't exist on BlockRun.

Fix: Detect native Anthropic model IDs (no / in the name) and replace them with the current model (blockrun/auto by default). Explicit BlockRun model IDs (provider/model format) pass through correctly.

Full Changelog: v0.9.3...v0.9.4

v0.9.3 — Expanded Model Catalog

What's New

50+ Models, 8 Providers

Massively expanded the model catalog with complete pricing and shortcuts.

New Model Shortcuts:

• codex → GPT-5.3 Codex | nano → GPT-5 Nano | o3 / o4 → OpenAI reasoning
• flash → Gemini 2.5 Flash | gemini-3 → Gemini 3.1 Pro
• grok-4 → Grok 4 | grok-fast → Grok 4.1 Fast Reasoning
• r1 → DeepSeek Reasoner | minimax → Minimax M2.7
• free / nemotron → Nemotron Ultra 253B | devstral / qwen-coder / maverick → NVIDIA free models

Updated Smart Routing

• Eco tier now uses Nemotron Ultra 253B (free, 253B params) as primary
• Better fallback chains with Minimax, Qwen3 Coder, Mistral Large
• Default free model upgraded from GPT-OSS 120B → Nemotron Ultra 253B

Complete Pricing

• All 50+ models now have accurate input/output token pricing
• Organized by provider: NVIDIA (free), Anthropic, OpenAI, Google, xAI, DeepSeek, Minimax, others

Full Changelog: v0.9.2...v0.9.3

v0.9.2

Fixes

• Streaming tool_use: forward delta.tool_calls through SSE stream (backend)
• DeepSeek-R1 reasoning: forward reasoning_content as thinking blocks (backend)
• Login page bug: wrap backend errors in Anthropic format so Claude Code shows the error instead of a login prompt
• max_tokens: always inject default (4096, adaptive) to prevent 400 on Turn 2+
• Token stats: search full SSE response for output_tokens, not just last chunk
• Streaming memory: cap accumulated response at 5MB
• Version: read from package.json at runtime
• Port validation: reject invalid ports (1-65535)
• Error messages: better errors for balance and models commands
• Model pricing: complete pricing table for all router-referenced models

Upgrade

npm update -g @blockrun/cc

v0.9.1

Fixes

• Always inject max_tokens default to prevent 400 errors on Turn 2+ conversations

CI

• Added build/typecheck CI workflow (Node 20 + 22)
• Added npm publish workflow with provenance