Terminal-first harness workflow
The product is designed around one terminal workflow for discovering, building, running, evaluating, and optimizing coding-agent harnesses, rather than moving between separate tools for each step.
SuperQode
SuperQode is an open source harness layer for coding agents that supports ACP, A2A, MCP, and UHP interoperability from one terminal. It helps developers discover, connect, evaluate, and optimize coding-agent harnesses, including a vendor-neutral A2A shortlist workflow.
What SuperQode is
SuperQode is an open source harness layer for coding agents. The site positions it as the layer around the agent you already use, with a terminal-first workflow for discovering, building, running, evaluating, and optimizing coding-agent harnesses.
It is built around interoperability across ACP, A2A, MCP, and UHP. The product site shows both a local coding-agent workflow and a published A2A agent that can return a vendor-neutral shortlist from the Harness Hub, while the Hub itself provides a catalogue of coding agents, ACP routes, runtimes, and model presets.
Core capabilities
Protocol interoperability
SuperQode states interoperability across ACP, A2A, MCP, and UHP, so the same layer can connect agents, harnesses, tools, and portable harness descriptions.
A2A shortlist agent
The A2A agent publishes an Agent Card and responds over JSON-RPC and HTTP+JSON. The site says it can return a ranked shortlist of third-party coding agents and harnesses from the Harness Hub.
Browseable harness catalogue
The Harness Hub shows 104 harnesses with status, runtime, setup, provenance, and licence information presented up front, including coding agents, ACP routes, runtimes, and model presets.
Harness evaluation and promotion flow
The site includes built-in harnesses, memory options, search routes, and optimization commands such as bench, eval, promote, and drift, with candidate artifacts kept reviewable before promotion.
Common use cases
Select a coding agent or harness
Use the Hub and terminal commands to compare coding agents, ACP routes, runtimes, and model presets before choosing a harness for a project.
Ask for an A2A shortlist
Run the A2A agent when another client or agent needs a vendor-neutral shortlist of coding agents without sending the repository.
Operate from one terminal
Use the same terminal session to connect an agent, browse the Hub, and call the A2A agent end to end when you want a single operator workflow.
Evaluate and promote harness changes
Use the benchmark, evaluation, promotion, and drift commands to test candidate harnesses against repository tasks before making them part of your workflow.
Run privately
Self-host the A2A server or run it in your own cloud when you need the agent inside your own network boundary.
Pros and Cons
Pros
- Open source and Apache 2.0, with self-hosting supported.
- Terminal-first workflow that combines discovery, connection, evaluation, and optimization in one place.
- Cross-protocol support across ACP, A2A, MCP, and UHP.
- The A2A agent can return a ranked shortlist from the public Hub without sending a repository.
- The Harness Hub presents provenance, licence, runtime, and setup information up front.
Cons
- The pricing page currently returns 404, so published pricing is not available from the site.
- Several catalogue entries and integrations are marked as pending or optional, so not every listed route appears to be fully complete.
FAQ
What does SuperQode do?
It provides a harness layer for coding agents. The site says you can discover, build, run, evaluate, and optimize coding-agent harnesses from one terminal, and that the product is interoperable over ACP, A2A, MCP, and UHP.
How do you use it?
The product pages describe terminal-first CLI and TUI workflows. Examples shown include installing with a shell script, connecting to an A2A endpoint, serving an A2A agent, and browsing the Harness Hub from the terminal.
Is it self-hosted or hosted?
Yes. The site says SuperQode is open source, Apache 2.0, and available as a self-hosted package. The A2A page also says the same server can be run privately in your own cloud.
What kinds of agents or harnesses are in the Hub?
The Hub lists 104 coding agent harnesses, including coding agents, ACP routes, runtimes, and model presets. It also shows open source, proprietary, and optional integration entries.
What does it cost?
No pricing details are available on the current pricing URL, which returns 404. The site does describe hosted, self-host, and private deployment options for the A2A agent, but not published price tiers.
Quick Facts
- Category
- Developer Tool
- Source domain
- superqode.dev
- License
- Apache 2.0
- Deployment
- Open source, self-hosted, hosted A2A endpoint, private cloud option
- Protocols
- A2A, ACP, MCP, UHP
- Hub size
- 104 harnesses
Alternatives à SuperQode
CreateOS Sandbox
CreateOS Sandbox is an isolated compute environment for running code and agent workloads inside Firecracker micro-VMs. It is designed for workflows that need machine-level isolation, private networking between sandboxes, and programmatic control through SDK, CLI, or MCP.
ByteAsk
ByteAsk is a terminal-first AI coding agent for C and C++ that edits repositories and verifies changes with the real compiler, debugger, sanitizers, and tests before showing a diff. It offers a free tier plus paid plans, with editor connectors and zero-retention handling described in the source.
hob
hob is an independent workspace for coding agents that keeps agent sessions, terminals, history, and follow-up work organized around the tools and providers you already use. It is aimed at developers who want local control over routing, history, and workspace structure rather than a bundled model stack.
Manta AI
Manta AI is an autonomous web app testing tool for teams that want to map application behavior, catch regressions, and generate tests without writing scripts or maintaining selectors. It works from a URL and supports plain-English test flows, run results with screenshots, and scheduled or deployment-triggered checks.
SonOf
SonOf connects to your repo and PM tool, audits the codebase and surrounding product context, and turns approved work into shipped tickets with senior engineering review. It is aimed at founders and engineering leaders who need backlog help without hiring a full team immediately.
Ghost
Ghost est un assistant IA en terminal pour discuter, générer du code et lancer des tâches en ligne de commande. Modèles gratuits, Linux, macOS, Windows, open source.