deepclaude: Claude Code Local Proxy with Live Backend Switching (DeepSeek / OpenRouter / Fireworks / Anthropic)

Score8.2

Popularity0.0

Riskconditional

TierGold

Score breakdown

Usefulness9.0

Novelty8.0

Momentum8.0

Maturity6.4

Open-source/build8.4

Evidence7.2

Workflow potential9.3

Setup ease8.8

Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.

Why it matters

Most Claude Code users today pay Anthropic prices ($3.00/M input + $15.00/M output for Opus, $200/month with usage caps) regardless of whether the task is routine file editing or hard reasoning. aattaran/deepclaude inverts that pattern: a single MIT localhost proxy for Claude Code that lets the consumer swap the model to DeepSeek / OpenRouter / Fireworks AI / or Anthropic with live mid-session swi

Who should use it

Claude Code users + AI coding agent developers + AI-curious readers tracking the coding-agent cost surface + engineering teams running Claude Code at scale + freelancers and indie developers paying for Claude Code out of pocket + any developer wanting to use Claude Code's autonomous agent loop while paying DeepSeek / OpenRouter / Fireworks prices instead of Anthropic pricesClaude Code users + tool-ergonomics users that want the localhost proxy architecture (the proxy intercepts all Claude Code API calls, sets per-session env vars, and restores the originals on exit) -- the right tool-ergonomics primitive for any Claude Code user who has been manually setting ANTHROPIC_BASE_URL before each sessionClaude Code users + live-switching users that want the live mid-session switching (slash commands inside an active session, no restart, no terminal commands) -- the right live-switching primitive for any Claude Code user who has been restarting Claude Code to switch backendsClaude Code users + backend-choice users that want the 4-backend support (DeepSeek / OpenRouter / Fireworks / Anthropic with per-backend pricing comparison) -- the right backend-choice primitive for any Claude Code user who has been locked into Anthropic pricingClaude Code users + cost-tracking users that want the `/_proxy/cost` endpoint (per-backend token usage + Anthropic-equivalent cost + savings percentage) -- the right cost-tracking primitive for any Claude Code user who has been paying Anthropic prices without visibility into savingsClaude Code users + remote-control users that want the optional `--remote` flag (Anthropic bridge WebSocket for remote control + localhost proxy for model calls) -- the right remote-control primitive for any Claude Code user who has been tethered to a desktop terminal

Who should skip it

Pass on deepclaude: Claude Code Local Proxy with Live Backend Switching (DeepSeek / OpenRouter / Fireworks / Anthropic) if its scope or audience does not match what your team is building right now.

About this signal

deepclaude: Claude Code Local Proxy with Live Backend Switching (DeepSeek / OpenRouter / Fireworks / Anthropic) is tracked by RepoRadar as a tool in the Radar section. It was first seen on 2026-07-08 and last updated on 2026-07-08. The current verdict is 'try now' with a Gold tier and easy setup difficulty. Across RepoRadar's eight signals, deepclaude: Claude Code Local Proxy with Live Backend Switching (DeepSeek / OpenRouter / Fireworks / Anthropic) is strongest on workflow potential (9.3) and practical usefulness (9.0) and weakest on maturity (6.4) — a profile worth weighing against your own priorities. This page summarizes the evidence RepoRadar has captured from captured source metadata. The score, tier, risk label, and verdict on this page are never influenced by sponsorship, ads, or tips — they reflect only the usefulness, popularity, novelty, momentum, maturity, and evidence signals described in the RepoRadar methodology.

How this item is evaluated

RepoRadar assigned deepclaude: Claude Code Local Proxy with Live Backend Switching (DeepSeek / OpenRouter / Fireworks / Anthropic) a composite score of 8.2 out of 10, placing it in the Gold tier. This score combines weighted sub-signals: usefulness (35%), novelty (18%), momentum (14%), maturity (10%), open-source/build quality (7%), evidence quality (6%), workflow potential (6%), and setup ease (4%). Popularity is tracked separately at 0.0 and never affects the composite score or tier. The risk label of 'conditional' reflects inherent user-impacting hazards, not generic novelty. Items with no risk flag may still require normal code review before production use.

Putting this into practice? Read How to evaluate an AI tool before you adopt it for the checklist behind this score.

Risk explanation

The 2193* / 147-fork / 14-subscriber repo is at active maintenance but the README references `DeepSeek V4 Pro` (a forward-looking / fictional model name per cycle 146 doctrine -- DeepSeek V4 Pro is not a publicly released DeepSeek model name as of cycle 369 verify time; the publicly released DeepSeek model family is V3 / V3.1 / V3.2); the proxy code does NOT depend on the V4 Pro model name and works against any Anthropic-compatible backend; so the cycle 146 risk_flag + conditional verdict pattern applies (NOT a strict reject) -- the consumer SHOULD treat the README's V4 Pro claim as forward-looking marketing and SHOULD verify the model name against the DeepSeek API documentation before adopting.

Evidence links

github.com

Closest alternatives / related signals

open-sourcemitaattarandeepclaudeclaude-codelocalhost-proxybackend-switchingdeepseek