cd /news/ai-agents/qode-the-first-ai-agent-that-can-gen… · home topics ai-agents article
[ARTICLE · art-30208] src=github.com ↗ pub= topic=ai-agents verified=true sentiment=↑ positive

Qode – The first AI agent that can generate 50k line codebases in one prompt

Qode, a terminal-based autonomous coding agent, has been released as an open-source Node SDK that can generate 50,000-line codebases from a single prompt. The tool features a four-phase Modus Maximus pipeline, multi-LLM support, and a verification system with self-correction, enabling developers to automate complex codebase operations and cross-cutting transformations.

read4 min views1 publishedJun 16, 2026

Node SDK v0.1.0

MIT

Qode is a terminal-based autonomous coding agent for long-running tasks and massive codebase development. It operates directly in your terminal with a rich TUI, connects to multiple LLM providers, and orchestrates multi-file changes through an intelligent pipeline that scales from one-line fixes to cross-cutting transformations spanning hundreds of files.

The Node SDK is a thin client that reuses the same Qode CLI configuration, tools, skills, and MCP servers. It streams responses in real time, surfaces approvals and tool calls, and lets you orchestrate sessions programmatically.

Qode provides an autonomous agent runtime, enabling you to:

Build custom applications— Integrate Qode into your own tools and workflows** Automate complex tasks**— Script multi-turn conversations across long-running sessions** Run massive codebase operations**— Use Modus Maximus for cross-cutting changes

Two execution modes: AUTO (single-agent turn loop with classifier-driven behavior) and MODUS MAXIMUS (4-phase pipeline: plan generation, user confirmation, sequential sub-agent execution through specialist profiles, and final summary)Intent classification: Heuristic analysis of every prompt for scope, depth, file references, action verbs, parallelism needs, and verification requirementsDynamic escalation: Runtime mode escalation via DynamicReclassifier when the current execution strategy is insufficient** Modus Maximus**: 15-50 dependency-aware plan steps, user confirmation (Looks Good / Needs Revision / Redo), sequential sub-agent execution with heuristic profile resolution, and final summary (Aims to generate 50k-80k lines of code for a fresh project in one shot through our specialized architecture)Specialist agent profiles: Editius (surgical code editing via StrReplace), Rewritius (full-file rewrites and refactoring), Searchius (codebase analysis), Auto (task-adaptive); switchable via/agent

Four-tier memory: WorkingMemory with priority-tagged compaction, EpisodicRecall with TF-IDF scoring, LTPM (disk-backed persistence with retention policies), SemanticRecall (vector-based ANN search via HNSW index), CodebaseGraph (language-aware model for TS, JS, Python, Rust, Go, Java)Verification pipeline: 7 gates (syntax, lint, typecheck, unit tests, integration tests, architecture, full suite) with per-language auto-detection and SHA-256 cachingSelf-correction: Automated fix-and-reverify loop with architecture escalation** Multiple LLM providers**: Anthropic, OpenAI, Google Gemini, Ollama (local, no API key), Kimi, OpenAI-compatible** Plugin system**: Manifest-based discovery with lifecycle hooks** Skill system**: Reusable prompt templates and inline scripts** MCP client**: stdio and HTTP/SSE transport** Session persistence**: JSONL wire format, blob store, migration system** Non-interactive mode**:--prompt

for CI/CD,--output-format json

Onboarding wizard: First-run setup for provider and model

Package Description Status
apps/q-cli
Main CLI — orchestrator, TUI, memory, MCP, plugins, verification Available
packages/agent-core
Core agent runtime — context, turn loop, tools, sub-agents, profiles Available
packages/qprovs
LLM provider abstraction Available
packages/qmain
Execution environment — file ops, shell, git, web Available
packages/node-sdk
Programmatic Node.js SDK Available
packages/oauth
OAuth 2.0 for MCP server auth Available
packages/telemetry
Optional crash reporting and telemetry Available
npm install -g qode-agent

Requires Node.js >= 22.19.0 and pnpm >= 10.33.0. The installer builds the project, creates ~/.Q/

, and installs the q-cli

wrapper.

q-cli

q

q-cli -p "Add error handling to src/routes/users.ts"

q-cli
q> /mode modus-maximus
q> Refactor the authentication system to use OAuth 2.0

q-cli -S <session-id>

q-cli -p "Fix the type errors" --output-format json
-S, --session <id>       Resume a session
-C, --continue           Continue last session
-y, --yolo               Auto-approve all actions
-m, --model <name>       Override LLM model
-p, --prompt <text>      Non-interactive mode
--plan                   Plan mode on startup
--auto                   Auto permission mode
--setup                  Re-run setup wizard
--output-format <fmt>    text | json | stream-json
--skills-dir <dir>       Additional skill directories
--cwd <path>             Working directory
--tui / --no-tui         Force/enable TUI
init          Initialize a Qode project
session       Manage sessions (list, show, delete, export, import)
config        View and edit configuration
doctor        Diagnose and fix configuration issues
migrate       Migrate data between versions
update        Check for and install updates
completions   Generate shell completion scripts
daemon        Start/stop the Qode daemon
connect       Connect to a remote instance
profile       Manage agent profiles
plugin        Manage plugins
/help             Interactive help dashboard
/status           Session status dashboard
/session          Show/manage session info
/clear            Clear transcript
/exit             Gracefully exit
/version          Show version info
/mode <name>      Switch mode (auto, modus-maximus)
/agent <name>     Switch profile (auto, editius, rewritius, searchius)
/qmd              Generate Q.md with project conventions
Q_PROVIDER    LLM provider (anthropic, openai, ollama, etc.)
Q_MODEL       Model name (e.g. claude-sonnet-4-20250514)
Q_API_KEY     API key for the provider
Q_BASE_URL    Custom base URL
Q_THINKING    Thinking level (none, low, medium, high)

Loaded in ascending priority: built-in defaults → ~/.Q/config.toml

.q/config.toml

(walked up from cwd) → environment variables. First run launches interactive onboarding wizard.

MIT

── more in #ai-agents 4 stories · sorted by recency
── more on @qode 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/qode-the-first-ai-ag…] indexed:0 read:4min 2026-06-16 ·