Repository-owned specification
Maintain context, memory, tools, workflows, policies, and evaluation criteria in a versioned HarnessSpec.
Agent Engineering
Terminal First
Open Source
SuperQode
Agent Engineering for Your Code Factory
Build your own coding-agent harnesses or connect the agents you already use. Orchestrate, evaluate, govern, and optimize how they work across your repositories.
install superqode
$
curl -fsSL https://super-agentic.ai/superqode.sh | shRequires Python 3.12 or higher. uv, uvx, and other install options
Build or connect
Evaluate
Govern
Optimize
Your Agent Engineering control plane
Build and own the system behind your coding agents
terminal first
Your harness
HarnessSpec: Harness as Code
Your models
local, BYOK, Open Models
Your memory
codebase context and adapters
Your rules
tools, policies, evaluation gates
Continuous learning loop
Turn failures and evaluation evidence into harness candidates, then promote only validated improvements.
Connect the agents you already use
Codex
Claude
Copilot
Antigravity
Grok
GLMPlus local models and BYOK providers. Providers docs
Open protocols
ACPexternal coding agents
MCPtools and resources
A2Aremote agent services
Harness ownership
One control plane for the agents working on your code
SuperQode keeps agent configuration, context, memory, tools, policies, workflows, and evaluation criteria in a versioned HarnessSpec maintained with your code.
Runtime independence
Use Codex, Claude, Copilot, Grok, OpenCode, local models, and compatible ACP agents with the same repository controls.
Controlled improvement
Evaluate harness candidates against repository tasks and promote changes only after the required gates pass.
The engineering problem
Coding agents do not create a dependable code factory by themselves
Every agent keeps its own context, tools, permissions, sessions, and evidence. SuperQode provides one engineering lifecycle across them.
01
Agent adoption is fragmented across products and repositories.
Current capability
Teams can choose specialized coding agents and models for different kinds of work.
Engineering gap
Each product maintains separate sessions, context, tools, permissions, and execution evidence.
02
The harness is rarely treated as an engineering artifact.
Current capability
Vendor-tuned loops can provide useful defaults and strong model-specific behavior.
Engineering gap
Teams still need to inspect, version, evaluate, and adapt the system around the model for their own repositories.
03
Agent output is not the same as verified repository delivery.
Current capability
Coding agents can plan, edit, test, and review increasingly complex changes.
Engineering gap
Orchestration alone does not provide isolation, recovery, policy enforcement, evidence, and an explicit delivery decision.
Agent Engineering layer
SuperQode code factory
A code factory is the organization-owned system of coding agents, harnesses, models, repositories, policies, evaluations, and delivery workflows that turns intent into verified code changes. SuperQode provides that system through a portable HarnessSpec, persistent sessions, durable WorkOrders, evidence, governance, and guarded optimization.
Use an agent you build, select a maintained harness, or connect an existing coding agent.
Read about harness independenceTerminal-first workbench
Run the complete Agent Engineering workflow from your terminal
Build or select a harness, connect any supported agent or model, switch persistent sessions, manage WorkOrders, and inspect evidence through the TUI and CLI.
Capabilities
One system for agent configuration, execution, and improvement
The terminal, specifications, protocols, delivery controls, and evaluation surfaces required to operate a code factory.
Interactive coding TUI
Switch models, pick harnesses, and inspect sessions without leaving the terminal.
HarnessSpec
A portable YAML contract for tools, permissions, model policy, memory, and workflows.
Durable WorkOrders
Dependency-aware tasks with isolated workers, recovery, evidence, and delivery gates.
Agent and model connections
Local runtimes, BYOK providers, ACP agents, and native vendor SDKs through one picker.
Evaluate and optimize
Score harness behavior, compare candidates, then promote or roll back under your gates.
Safety and policy
Approvals, sandbox behavior, trust checks, and plugin policy stay explicit and inspectable.
Local and open models
Run Ollama, MLX, LM Studio, vLLM, SGLang, TGI, and llama.cpp, including fully offline.
Sessions and continuity
Persistent sessions with lineage, harness switching mid-session, forks, handoffs, and export.
Start here
Install, open a repository, connect an agent
Everything after this runs from the TUI or the CLI. The short alias sq runs every superqode command.
terminal
curl -fsSL https://super-agentic.ai/superqode.sh | sh cd /path/to/your/project superqode # inside the TUI: :connect
Connect local models, BYOK providers, ACP agents, and native vendor runtimes. See the Connections docs.
Omnigent agent specifications can be imported as portable HarnessSpecs. See the Omnigent compatibility docs.
Maintained harness presets
SuperQode
KimiGLM
Qwen
DS4
Gemma
MiniMaxEvery install ships maintained presets for these model families. Harness system docs
Documentation
Full reference for every command and option
This page is an overview. The documentation covers installation, provider setup, the HarnessSpec schema, evaluation workflows, and the complete CLI.
Partner program
Design partners
Teams standardizing coding-agent workflows across repositories can join a limited design partner program. Partners evaluate SuperQode on real engineering workflows and help refine the product for production teams.
Agent Engineering for your code factory
Build or connect a coding-agent harness, then evaluate, govern, and improve how it works across your repositories.
Open source · Terminal first · Any agent or model · Local or cloud
Design partnersFAQ
Frequently asked questions
Core definitions, supported connections, and installation guidance.