AI Tools

DeepClaude Enables Claude Code Loop with DeepSeek V4 Pro, 17x Cheaper

DeepClaude lets users run Claude Code's autonomous agent loop using DeepSeek V4 Pro or other backends, cutting costs by 17 times compared to the original $200 monthly plan. It keeps the same terminal interface and features like file editing and bash execution while swapping the model backend. Setup takes two minutes with an API key, and it supports OpenRouter, Fireworks AI, and Anthropic options.

Neura News

Neura News

Neura Market Editorial

May 4, 20264 min read
DeepClaude Enables Claude Code Loop with DeepSeek V4 Pro, 17x Cheaper

DeepClaude Enables Claude Code Loop with DeepSeek V4 Pro, 17x Cheaper

Developers now have access to DeepClaude, a tool that runs Claude Code's autonomous coding agent loop powered by DeepSeek V4 Pro, OpenRouter, or any backend compatible with Anthropic's API. This setup delivers the identical user experience at a fraction of the cost, specifically 17 times cheaper than the standard Claude Code subscription.

Claude Code stands out as a top autonomous coding agent, but its $200 per month price includes usage limits. DeepSeek V4 Pro achieves a 96.4% score on LiveCodeBench while charging just $0.87 per million output tokens. DeepClaude replaces the underlying model, or brain, yet retains the full structure, or body, of the tool.

The terminal setup remains the same: Claude Code CLI handles tool loops, file editing, bash commands, and git operations without changes. API requests now route to DeepSeek V4 Pro at $0.87 per million tokens instead of Anthropic's $15 per million. All core functions operate normally, including file access, modifications, bash runs, subagent creation, and multi-step coding sequences. The sole variation lies in the model performing the reasoning.

Quick Setup in Two Minutes

Starting with DeepClaude requires a few simple steps. First, obtain a DeepSeek API key by signing up at platform.deepseek.com, adding $5 in credit, and copying the key.

Next, configure environment variables. On Windows using PowerShell, run: setx DEEPSEEK_API_KEY "sk-your-key-here". For macOS or Linux, execute: echo 'export DEEPSEEK_API_KEY="sk-your-key-here"' >> ~/.bashrc followed by source ~/.bashrc.

Installation varies by platform. On Windows, copy the script to a PATH directory like $env:USERPROFILE.local\bin\deepclaude.ps1, or add the repository folder to PATH with setx PATH "$env:PATH;C:\path\to\deepclaude". On macOS or Linux, make the script executable with chmod +x deepclaude.sh, then create a symlink: sudo ln -s "$(pwd)/deepclaude.sh" /usr/local/bin/deepclaude.

Usage is straightforward. Launch with deepclaude to start Claude Code using DeepSeek V4 Pro. Check status with deepclaude , status for backends and keys. Switch providers via deepclaude , backend or for OpenRouter at $0.44 per million input tokens, deepclaude , backend fw for Fireworks AI, or deepclaude , backend anthropic for the original Claude. View costs with deepclaude , cost or test latency with deepclaude , benchmark.

How It Operates and Backend Support

DeepClaude configures environment variables temporarily for each session to redirect API calls. It reads variables like ANTHROPIC_BASE_URL for the endpoint (default api.anthropic.com), ANTHROPIC_AUTH_TOKEN for authentication, ANTHROPIC_DEFAULT_OPUS_MODEL for high-end tasks, ANTHROPIC_DEFAULT_SONNET_MODEL for mid-tier, ANTHROPIC_DEFAULT_HAIKU_MODEL for subagents, and CLAUDE_CODE_SUBAGENT_MODEL for spawned agents. Upon exit, it resets to original settings.

The #1 Newsletter in AI

Stay ahead of the AI curve

The most important updates, news, and content — delivered weekly.

No spam. Unsubscribe anytime.

Supported backends include DeepSeek (default, , backend ds, $0.44 input/$0.87 output per million, China servers with auto context caching), OpenRouter (, backend or, same pricing, US servers for low latency), Fireworks AI (, backend fw, $1.74 input/$3.48 output, fastest US inference), and Anthropic (, backend anthropic, $3 input/$15 output, original Opus for tough tasks).

Set keys as needed: DEEPSEEK_API_KEY for DeepSeek, OPENROUTER_API_KEY for OpenRouter, FIREWORKS_API_KEY for Fireworks.

Pricing and Feature Details

Cost savings are significant. Light use over 10 days monthly hits Anthropic's $200 cap, but DeepClaude costs about $20, a 90% reduction. Heavy use over 25 days caps at $200 versus $50, saving 75%. With auto loops, it's $200 versus $80, or 60% less. DeepSeek's caching drops repeat context to $0.004 per million from $0.44 uncached.

Most features work fully: file read/write/edit, bash/PowerShell, glob/grep search, tool loops, subagents, git, project init, and thinking mode. Limitations include no image input due to backend support, sequential tool execution instead of parallel, no MCP server tools, ignored Anthropic cache_control with DeepSeek's own caching instead. For routine tasks covering 80% of work, DeepSeek V4 Pro matches Claude Opus; for complex 20%, use , backend anthropic.

IDE and Remote Integration

Integrate with VS Code or Cursor by adding terminal profiles. On Windows, edit settings JSON with {"terminal.integrated.profiles.windows": {"DeepSeek Agent": {"path": "powershell.exe", "args": ["-ExecutionPolicy", "Bypass", "-NoExit", "-File", "C:\path\to\deepclaude.ps1"]}}}. On macOS/Linux: {"terminal.integrated.profiles.linux": {"DeepSeek Agent": {"path": "/usr/local/bin/deepclaude"}}}.

Remote control opens sessions in browsers: deepclaude , remote for DeepSeek, with options for other backends. It generates a claude.ai/code/session_... URL. Requires Claude auth login, claude.ai subscription, and Node.js 18+. A local proxy splits WebSocket to Anthropic's bridge and model calls to the chosen backend.

The project uses an MIT license. DeepSeek directory, based in China, produces cost-effective models like V4 Pro that rival top performers on benchmarks. Anthropic's Claude series powers advanced agents, with Claude Code focusing on terminal-based autonomy.

Related on Neura Market

More from Neura News

Funding

Prentis AI Lab Co-Founded by Reid Hoffman, Marc Pincus Seeks $100M

Prentis, a new AI research lab co-founded by Ritankar Das, Reid Hoffman, and Marc Pincus, is in talks to raise $100 million at a $1 billion valuation. The startup focuses on computer use models that automate office workflows. It has already signed contracts worth up to $50 million with several customers and claims its Hive-32B model outperforms rivals like OpenAI's GPT-5.4 and Anthropic's Claude Opus 4.6 on key benchmarks.

Jul 24·4 min read
Industry

Cognition Acquires Poke to Give Devin Coding Agent a Personality

Cognition, the startup behind AI coding assistant Devin, has acquired Poke, an AI assistant known for its friendly, conversational style. The deal, valued in the low nine figures, aims to bring Poke's personality-driven interaction model to Devin, making the coding agent feel more like a colleague than a tool. Poke will also benefit from Cognition's models and infrastructure to become faster and more reliable.

Jul 24·3 min read
AI Models

Anthropic expands Claude voice mode to Opus and Sonnet models

Anthropic has expanded Claude's voice mode to run on its most powerful models, Opus and Sonnet, across mobile, desktop, and web platforms. Users can now switch between models mid-conversation, use voice commands in eleven languages, and connect to external tools like Gmail, Google Calendar, or Slack to compose and send emails by voice. The update positions Claude as a unique option for tool integration in voice AI, though competitors like OpenAI and Google offer more natural speech processing.

Jul 24·2 min read