What our run established
Our run took place in our sandbox on October 1, 2026, at commit fce5fa3. The stored installation outcome is success, exit code 0, with an 11-second rounded duration. The raw record retains 11,340 milliseconds. These are the attached run's values, rather than a new execution.
The installation tail contains an npm ci EUSAGE error explaining that a lockfile is required. Later in the same tail, postinstall reports downloading the md2wechat 3.8.0 Linux binary, followed by an up-to-date message. Read the successful overall outcome alongside that earlier error; it does not establish that npm ci itself succeeded.
Our harness skipped build and test because its selected npm path found no corresponding scripts. The detector also recorded Go and a Makefile. The skipped npm steps therefore do not establish that the repository lacks build instructions or tests. We did not verify a Go source build, execute a rendering workload, generate an image, upload an asset or create a draft.
The stored dependency summary reports count 0 and 30 MB on disk. Its npm audit summary records 0 advisories. Scanned-package coverage is unknown because the attached audit record has no scannedPackages field. That advisory result is not evidence that the downloaded binary or remote services are secure.
The testing-methodology page explains how MrKeyoor records command outcomes, timings and limitations. It provides provenance context; its current description does not establish which methodology version governed this older run.
Free AI mode leaves HTML generation unfinished
md2wechat-skill: free AI mode hands HTML work to an external model.
The supplied README distinguishes two output paths. Free AI mode generates a prompt for a host agent or external language model to continue. Successful API conversion returns final WeChat HTML and requires an API key. Choose between them according to where you want HTML generation to happen, rather than assuming the free path produces the same finished artifact.
The documented command boundaries help separate inspection from publishing. Inspection reports metadata, checks, readiness targets and blockers. Preview writes successful API conversion output to a local file. Conversion performs upload or draft effects only when explicitly requested. These are documented capabilities, not outcomes verified by our sandbox.
The README also says local layout validation checks syntax and cannot prove that the remote renderer has deployed a feature. Its newer interactive modules default to a complete static display, while their explicit first-layer behavior remains unverified inside WeChat. Preview a representative article and inspect the result in the intended publishing environment before relying on those effects.
Publishing requires credentials and an allowed network exit
The credential guide identifies AppID and AppSecret as requirements for image uploads, drafts and image messages. It also explains that many WeChat interfaces check the public IP of the machine making the request. A correctly configured account can still receive a whitelist error when the request comes from another network exit.
For a workstation, server or CI job, establish which machine actually sends the WeChat request before configuring the allowlist. The guide offers the paid API's fixed-egress option for environments without a stable public IP. That proxy applies to WeChat operations; the guide says it does not govern rendering or image-provider traffic.
A useful next step is to follow the documented sequence: validate configuration, try a single image upload, then test draft creation before attempting the complete publishing chain. Those are proposed checks for your environment, not checks completed in our run.
API access and commercial permission are separate decisions
The license requires separate written authorization for revenue-related internal use within a for-profit organization and for services delivered to paying clients.
The license grants specified personal non-commercial production uses, including personal content creation without direct commercial compensation for that content. It expressly distinguishes passive platform advertising revenue from direct compensation. It also permits the listed learning, evaluation, nonprofit and contribution uses.
The current README says buying API service does not automatically include commercial authorization. Resolve both questions before adopting it for a company or client workflow. The license lists January 1, 2030 as its Change Date and also includes an earlier fourth-anniversary provision for a specific version. Read the version's applicable terms rather than treating a future Apache license as permission for present commercial use.
Alternatives depend on how you want to publish
Choose WeChat Markdown Editor when you want browser editing with immediate Markdown-to-WeChat rendering.
Its README documents custom CSS, local drafts, file import and export, and several deployment options. That makes it a formatting-workflow alternative for readers who prefer an editor. The supplied extract does not establish it as a replacement for md2wechat's agent discovery and guarded draft commands.
Markdown Nice is another editor option: its README documents custom styles and support for WeChat, Zhihu and Juejin. Compare it when the decision is where to compose and format an article. Its supplied extract does not establish equivalent publishing automation.
The wechat-format README explicitly says the project is discontinued. It documents Markdown-to-WeChat HTML conversion, but should not remain an unqualified recommendation for a new workflow.
How to read the ratings
The rating scale is 1 for substantial documented obstacles, 2 for considerable manual work, 3 for a workable path with material prerequisites, 4 for clear supporting documentation or demonstrated capability, and 5 for comprehensive verification within the assessed dimension. An unassessed dimension has no numerical rating.
Setup receives 3 because installation completed but publishing requires service and account configuration. Documentation receives 4 because the supplied README and credential guide explain command boundaries, configuration and troubleshooting in detail; this assessment concerns those Chinese materials. Maturity receives 3 because the documented workflow has explicit boundaries, while our run verified installation only. Community is unassessed because the supplied current extracts do not establish contributor participation, response times or maintenance history.