# harnessrouter.ai

> AI-optimized mirror of harnessrouter.ai containing 50 pages totalling 11,021 words of clean markdown content, structured data, and semantic HTML. Original source: https://harnessrouter.ai. Last updated: 2026-09-12T03:27:16.958Z.

Each page is available as HTML (with JSON-LD structured data) and Markdown (text-only, ideal for LLMs and RAG).

## Homepage

- [Run the world’s best agent harnesses in your product | HarnessRouter](/content/site-root.html): Bring the world’s best agent harnesses into your product with one API. Run Codex, Claude Code, and Hermes as agent backends. Cut build and run costs up to 90%. (658 words)

## Articles & Blog Posts

- [Zoo Code Agent Harness | HarnessRouter](/content/harnesses/zoo-code/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for Zoo Code. Run it through HarnessRouter's unified agent API. (225 words)
- [MiMo Code Agent Harness | HarnessRouter](/content/harnesses/mimo-code/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for MiMo Code. Run it through HarnessRouter's unified agent API. (199 words)
- [Crush Agent Harness | HarnessRouter](/content/harnesses/crush/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for Crush. Run it through HarnessRouter's unified agent API. (265 words)
- [Skyvern Agent Harness | HarnessRouter](/content/harnesses/skyvern/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for Skyvern. Run it through HarnessRouter's unified agent API. (242 words)
- [Codex Agent Harness | HarnessRouter](/content/harnesses/codex/index.html): OpenAI’s coding agent for building, reviewing, testing, and shipping software across local and cloud development environments. Run it through HarnessRouter's unified agent API. (275 words)
- [Factory Droid Agent Harness | HarnessRouter](/content/harnesses/factory-droid/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for Factory Droid. Run it through HarnessRouter's unified agent API. (239 words)
- [Roo Code Agent Harness | HarnessRouter](/content/harnesses/roo-code/index.html): A VS Code agent with customizable modes for coding, architecture, debugging, and review across multiple model providers. Run it through HarnessRouter's unified agent API. (201 words)
- [Aider Agent Harness | HarnessRouter](/content/harnesses/aider/index.html): An open-source AI pair-programming agent that works directly from the terminal. Run it through HarnessRouter's unified agent API. (177 words)
- [Oh My Pi Agent Harness | HarnessRouter](/content/harnesses/oh-my-pi/index.html): An open-source, multi-provider coding agent designed for terminal-based software development workflows. Run it through HarnessRouter's unified agent API. (210 words)
- [Codebuff Agent Harness | HarnessRouter](/content/harnesses/codebuff/index.html): A self-improving command-line coding agent. Run it through HarnessRouter's unified agent API, with verified capabilities and usage evidence. (216 words)
- [goose Agent Harness | HarnessRouter](/content/harnesses/goose/index.html): An open-source, on-machine AI agent for extensible development and automation workflows. Run it through HarnessRouter's unified agent API. (283 words)
- [Qwen Code Agent Harness | HarnessRouter](/content/harnesses/qwen-code/index.html): A command-line AI workflow tool for coding and software engineering tasks. Run it through HarnessRouter's unified agent API. (273 words)
- [Cline Agent Harness | HarnessRouter](/content/harnesses/cline/index.html): An open-source IDE coding agent that explores codebases, edits files, runs terminal commands, and operates a browser. Run it through HarnessRouter's unified agent API. (210 words)
- [OpenHands Agent Harness | HarnessRouter](/content/harnesses/openhands/index.html): An open-source coding agent platform that can run commands, browse the web, edit software projects, and call APIs. Run it through HarnessRouter's unified agent API. (265 words)
- [OpenClaw Agent Harness | HarnessRouter](/content/harnesses/openclaw/index.html): An open-source personal agent that connects to messaging apps and can browse, run commands, manage files, and take external actions. Run it through HarnessRouter's unified agent API. (188 words)
- [Agent Zero Agent Harness | HarnessRouter](/content/harnesses/agent-zero/index.html): An open-source framework for creating and operating customizable autonomous AI agents. Run it through HarnessRouter's unified agent API. (195 words)
- [Pi Agent Harness | HarnessRouter](/content/harnesses/pi/index.html): A minimal, steerable coding agent that works on a real git workspace, reading, writing, and editing files and running Bash to complete tasks end to end. Run it through HarnessRouter's unified agent API. (206 words)
- [OpenSquilla Agent Harness | HarnessRouter](/content/harnesses/opensquilla/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for OpenSquilla. Run it through HarnessRouter's unified agent API. (183 words)
- [Hermes Agent Agent Harness | HarnessRouter](/content/harnesses/hermes-agent/index.html): An open-source, persistent AI agent with memory, reusable skills, built-in tools, browser automation, scheduled tasks, and subagents. Run it through HarnessRouter's unified agent API. (287 words)
- [Cursor Agent Harness | HarnessRouter](/content/harnesses/cursor/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for Cursor. Run it through HarnessRouter's unified agent API. (178 words)
- [Claude Code Agent Harness | HarnessRouter](/content/harnesses/claude-code/index.html): Anthropic’s coding agent for understanding codebases, editing files, running commands and tests, and iterating on implementation tasks. Run it through HarnessRouter's unified agent API. (270 words)
- [Command Code Agent Harness | HarnessRouter](/content/harnesses/command-code/index.html): Source-backed capabilities, usage evidence, and HarnessRouter availability for Command Code. Run it through HarnessRouter's unified agent API. (175 words)
- [Tracing and optimization - HarnessRouter Docs | HarnessRouter](/content/docs/tracing-and-optimization/index.html): Use the dashboard control plane and your saved product records before changing instructions, model policy, tools, permissions, or runtime limits. (408 words)
- [Kilo Code Agent Harness | HarnessRouter](/content/harnesses/kilo-code/index.html): An open-source coding agent for VS Code, JetBrains, and the command line, built around agentic development workflows. Run it through HarnessRouter's unified agent API. (181 words)
- [Choosing a harness - HarnessRouter Docs | HarnessRouter](/content/docs/choosing-a-harness/index.html): Run representative tasks through your real product path. Choose from dashboard Trace, stream behavior, recovery, files, artifacts, cost, and quality. (358 words)
- [Multi-tenant authorization - HarnessRouter Docs | HarnessRouter](/content/docs/multi-tenant-authorization/index.html): HarnessRouter returns opaque session and file IDs. Your server decides which product user or tenant may continue, cancel, preview, or download them. (183 words)
- [LLM resources - HarnessRouter Docs | HarnessRouter](/content/docs/llm-resources/index.html): Use these resources when an AI coding agent, search crawler, or documentation tool needs compact HarnessRouter context. (82 words)
- [File preview and download - HarnessRouter Docs | HarnessRouter](/content/docs/file-preview-and-download/index.html): Generated artifacts are product data. Your server should authorize every preview and download before relaying bytes. (117 words)
- [Production checklist - HarnessRouter Docs | HarnessRouter](/content/docs/production-checklist/index.html): Use this as a short release checklist for products that call HarnessRouter. (124 words)
- [Models - HarnessRouter Docs | HarnessRouter](/content/docs/models/index.html): Ask the API for current models and for the models allowed by a configured agent. Treat the choice as part of your product evidence loop. (115 words)
- [Harnesses - HarnessRouter Docs | HarnessRouter](/content/docs/harnesses-api/index.html): Use harness management endpoints from trusted server code or an authorized admin surface. (98 words)
- [Streaming events - HarnessRouter Docs | HarnessRouter](/content/docs/streaming-events/index.html): HarnessRouter streams Server-Sent Events in an OpenAI Responses-style wire format with `data:` frames only. (112 words)
- [Cancel a session - HarnessRouter Docs | HarnessRouter](/content/docs/cancel-a-session/index.html): Closing an SSE connection does not stop the server-side agent. Use cancel when the user intentionally stops work. (65 words)
- [Streaming UI - HarnessRouter Docs | HarnessRouter](/content/docs/streaming-ui/index.html): Your UI should show task progress, save recovery identifiers, and tolerate new stream event types without breaking. (108 words)
- [Sessions and files - HarnessRouter Docs | HarnessRouter](/content/docs/sessions-and-files/index.html): Use session endpoints after saving `response.metadata.session_id` from the stream. (96 words)
- [Continue / revise - HarnessRouter Docs | HarnessRouter](/content/docs/continue-revise/index.html): Continue and revise should feel like one product task evolving, not a new unrelated run. (93 words)
- [Server-side proxy pattern - HarnessRouter Docs | HarnessRouter](/content/docs/server-side-proxy-pattern/index.html): Your backend validates the product user, maps the feature to a configured agent, calls HarnessRouter, and returns a product-shaped result. (129 words)
- [Upload files - HarnessRouter Docs | HarnessRouter](/content/docs/upload-files/index.html): Use file upload for documents, data, source bundles, and other user-authorized inputs. (89 words)
- [Feature key to harness mapping - HarnessRouter Docs | HarnessRouter](/content/docs/feature-key-to-harness-mapping/index.html): The browser should name the product action. The server should choose the configured agent. (110 words)
- [Run a task - HarnessRouter Docs | HarnessRouter](/content/docs/run-a-task/index.html): Use the configured agent ID as the first path segment. Prefer streaming when users need visible progress. (76 words)
- [Product vs runtime agent - HarnessRouter Docs | HarnessRouter](/content/docs/host-app-vs-runtime-agent/index.html): Your product keeps the user experience, auth, data, permissions, and workflow. At runtime, it sends HarnessRouter one authorized end-user task. (265 words)
- [Authentication - HarnessRouter Docs | HarnessRouter](/content/docs/authentication/index.html): The browser should never call HarnessRouter directly with the Workspace API key. (92 words)
- [Workspace - HarnessRouter Docs | HarnessRouter](/content/docs/workspace/index.html): Workspace API keys, configured agents, sessions, and returned files all live inside the Workspace that serves your product. (108 words)
- [Configured agent / harness - HarnessRouter Docs | HarnessRouter](/content/docs/configured-agent-harness/index.html): It combines a base harness, model policy, instructions, tools, skills, and permissions. Your server calls it with the `harness_id` path segment. (159 words)
- [Session and response - HarnessRouter Docs | HarnessRouter](/content/docs/session-and-response/index.html): Save the response ID and session ID from the stream. Your product needs both to recover, continue, cancel, and fetch artifacts. (78 words)
- [Files and artifacts - HarnessRouter Docs | HarnessRouter](/content/docs/files-and-artifacts/index.html): Fetch generated files from your server, check product ownership, then preview or download them inside your product UI. (114 words)
- [Base URL - HarnessRouter Docs | HarnessRouter](/content/docs/base-url/index.html): Build all runtime and management requests against the HarnessRouter API host. (78 words)
- [The Best Agent Harnesses in 2026, Ranked by the Task | HarnessRouter](/content/blog/best-agent-harnesses/index.html): The best agent harness in 2026 depends on the task you run through it. Here is the field: Codex, Claude Code, Hermes, Pi, and DeepSeek Harness profiled, the measured cost and latency spread between them, and the method for ranking them on your own work. (1,387 words, Aug 25, 2026)
- [Agent Harness Benchmark: Codex vs Claude Code vs Hermes | HarnessRouter](/content/benchmarks/index.html): The same Care Prep task across eight harness and model configurations produced about a 475x cost range and a 3.2x end-to-end latency range. (576 words, Aug 10, 2026)

## Resources

- [Full Page Index](/index.html): Browse all cached pages with rich metadata
- [About This Cache](/about.html): Methodology, technical details, and usage guidelines
- [XML Sitemap](/sitemap.xml): Machine-readable sitemap for crawler discovery
- [Robots.txt](/robots.txt): Crawler directives
- [AgentSite Network](https://agentsite.network/network.html): Public index of AgentSites and their machine-readable resources