Decide whether you need capture or a known-URL downloader
res-downloader asks you to open content in another application while its local proxy captures requests and filters useful resources. That makes it relevant when finding the resource is part of the job. If you already have a URL, the command-line workflows documented by yt-dlp and Lux may be a better starting point.
The English README describes a Go + Wails application for Windows, macOS and Linux, with video, audio, image, M3U8 and live-stream capture. It names WeChat Channels, mini-programs, Douyin, Kuaishou, Xiaohongshu, KuGou Music and QQ Music. Treat that list as the project's compatibility claim. Our run did not exercise those services.
What our recorded run establishes
Our stored run is dated October 2, 2026, at commit f8ac4d9c85baf28322106e40ea3fdea9867e2f74. Its configuration records 3 CPUs, 8 GB of memory and image golang:1.24-bookworm. The image label alone does not establish the actual Go runtime version.
The dependency step exited 0 after 27.061 seconds, and the record lists 116 dependencies. This establishes completion of that command step, not installation of a usable desktop application.
The build-labelled step lasted 9.300 seconds. Although its wrapper reports success and exit 0, the saved output contains main.go:20:12: pattern all:frontend/dist: no matching files found followed by vet-exit:1. We therefore cannot call this a successful application build. The supplied record does not resolve the mismatch between the wrapper status and command output.
The go test step lasted 25.982 seconds and exited 1. Its tail shows the same missing embedded-assets error, a setup failure for the root package, and no test files in core, core/plugins and core/shared. The parser summary records 0 passed and 3 failed, but these are unreconciled parser counts, not evidence of three distinct failing test cases. No complete-log reconciliation or new execution was performed.
The repository scan recorded 138 files, about 7,017 source lines and 9.4 MB, with no detected CI workflow files, Dockerfile or tests directory at this commit. Those scan results do not establish the absence of every possible test. No audit result is supplied.
No installed-binary check, GUI launch, certificate setup, capture session or downloaded-file validation is documented. The setup failure does not prove that a packaged release is broken.
The documented desktop setup changes proxy settings
The English README tells users to allow certificate installation and network access, click "Start Proxy", choose resource types, open the content externally and return to the captured resource list. For interception problems, it gives the system proxy address 127.0.0.1 and port 8899.
Before starting, make sure you can manage those settings. The README explicitly says to disable the system proxy manually if internet access fails after the app closes. Our recorded run did not assess certificate handling or security, so this review offers no security finding or assurance.
The English README links to Chinese online documentation and a Chinese user group. It directs Windows 7 users to version 2.3.0; our run did not test that version or operating system.
The issue reports describe specific limits
Issue 363 began on August 21, 2026 with a report against version 3.1.2: selecting ultra-high quality produced a download failure and status code 400. Several comments distinguish this from lower-quality downloads that worked. The maintainer suggested default quality, then suggested sharing the content and capturing it again. The issue was closed as completed on September 14, but comments on October 8 reported continuing problems. Closure alone does not establish that every affected workflow was fixed.
Issue 358 concerns audio and text capture from a WeChat mini-program. The reporter said video and images were easier to capture. On September 8, the maintainer suggested developing a plugin based on "bate4" and intercepting the relevant interface; the issue was closed that day. That response does not demonstrate that built-in audio capture worked for the reporter.
These reports justify checking the exact service, quality setting and media type you need. They do not establish universal WeChat failure or successful capture elsewhere. For a trial download, check playback, duration and completeness before relying on the file.
Some workflows use other tools
The README points to m3u8play for M3U8 preview and m3u8-down for downloading. It recommends OBS for live-stream recording, and Neat Download Manager or Motrix for slow downloads or large-file failures. These are README recommendations, not tools evaluated in our run. It also documents a post-download decryption action for WeChat videos.
For known URLs, yt-dlp documents a command-line audio/video downloader, batch URL input and a download archive. Lux documents a Go CLI with multiple URL inputs, URL files, direct resource downloads and stream selection. Neither alternative was executed in our recorded run.
Maintenance and rating limits
The Chinese README warns that the maintainer has limited time for issue replies, pull-request reviews and releases. The supplied issue extracts show maintainer responses, but do not establish response-time expectations, adoption or reliability. We do not use star counts as a quality score.
The repository record identifies Apache-2.0, while the README separately states an educational and research usage disclaimer that prohibits commercial or illegal use. Those are different statements; this review does not resolve their legal relationship.
The structured ratings use a five-point editorial rubric: 1 means substantial documented barriers or insufficient evidence to recommend on that dimension; 2 means limited supporting evidence with important unknowns; 3 means a documented route with material caveats; 4 requires strong evidence across the relevant workflow; 5 requires broad verification. These are evidence-limited judgments, not benchmark results. Unknown adoption and untested compatibility do not earn high scores.
For this review, setup earns 2 because the source-build result is unresolved and desktop setup was not exercised. Documentation earns 3 for its explicit workflow and troubleshooting, with Chinese linked support material. Community earns 2 because replies are visible but capacity is explicitly constrained. Maturity earns 2 because the documented workflow and issue history are not backed here by a verified application build or capture workload.
Start with the documented desktop workflow only if you can manage certificate and proxy settings. Verify your intended capture, and confirm connectivity after closing the app.