# tigicion/dao-code

Open-source TypeScript terminal coding agent for DeepSeek-V4 — builds on DeepSeek's strong price-performance and ultra-cheap cache pricing, engineering byte-stable prefixes and cache-reusing forks so cross-session memory and a continuous self-correction layer add almost no token cost; 1M context, Skills/MCP/Hooks, Claude Code config compatible.

Repository: https://github.com/tigicion/dao-code
Canonical: https://ross.abutalabs.com/products/dao-code
Homepage: https://www.npmjs.com/package/dao-code
Language: TypeScript
License: MIT
License Family: permissive
Topics: ai-agent, cli, coding-agent, deepseek, llm, terminal, typescript
Last push: 2026-08-02T17:43:03+00:00

## Health v2 (maintenance only)
Score: 76/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 95, release rhythm 93, longevity 6
- inputs: {"age_days": 86, "days_push": 31, "days_rel": 46, "gap_med": 3.5, "n_releases_24m": 9}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1254, forks 52 (observed 2026-08-28T04:04:08.811352+00:00)

## What it is
Dao Code is an open-source terminal-native AI coding agent written in TypeScript, targeting DeepSeek V4 with 1M context and Claude Code-compatible configuration. It engineers byte-stable prefixes and cache-reusing forks so cross-session memory and self-correction add almost no token cost, achieving ~95% prompt-cache hit rates.

## Use cases
- run an AI coding agent in the terminal with DeepSeek instead of Claude
- reduce LLM coding agent costs via prompt cache optimization
- get a Claude Code alternative that works in mainland China
- let an AI agent read, edit, and test code with approval gates
- maintain cross-session memory for a coding assistant cheaply
- use MCP tools and skills with a terminal coding agent

## When to choose
- you want a low-cost, pay-as-you-go coding agent without Anthropic access or quota limits
- you work in a terminal and want streaming reasoning with an approval gate
- you need 1M-context coding assistance with persistent memory at minimal token cost
- you already use Claude Code config and want a compatible DeepSeek-based agent

## When to avoid
- you need a GUI or IDE-integrated coding assistant
- you require top-tier closed-model quality regardless of cost
- you don't want to use DeepSeek models specifically
- the project is young (v0.4.x) and you need battle-tested stability

## Facets
- artifact type: cli-tool
- maturity: active
- function: agent-framework, llm-inference, cli, terminal-ui, prompt-engineering, mcp, chatbot
- domain: developer-tools, artificial-intelligence, large-language-models
- platform: cli, cross-platform, windows
- tags: coding-agent, deepseek, prompt-caching, tui, claude-code-alternative, cost-optimization, ink, command-line, ai-agents, nodejs, macos, linux

## Member repositories
- tigicion/dao-code (main) score 76

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:08.811352+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:07:25.907680+00:00, confidence not recorded.
  - readme: https://github.com/tigicion/dao-code (fetched 2026-08-28T04:04:08.811352+00:00, sha 523fa7fb8732)
  - registry_npm: https://registry.npmjs.org/dao-code (fetched 2026-08-29T12:18:05.806193+00:00, sha 00eae06b19cb)
- Data as of 2026-08-30T08:39:29.467469+00:00.