使用说明
更新于 2026-10-04 · 适用于文脉在线试用版
哪些功能可以直接用
- 样例:免注册,附预先准备的文章分析和问答,并非实时生成。
- 导入与阅读:无需账户。支持粘贴文字、TXT、Markdown、文字版 PDF 和部分公开网址。
- 在线 AI:需登录。可用平台试用额度,无需配置模型;也可主动使用自己的智谱 API Key。两种方式都由文脉服务器调用模型。
试用多久、能用几次
新账户的平台试用目前有 7 天、24 次 AI 请求,以账户页显示为准。一次操作可能包含多次请求,不等于能问 24 个问题。到期或用完后,平台试用请求停止,已保存的文章和问答仍能查看。目前没有付费续用入口。
也可在设置 → 模型服务保存并启用自己的智谱 API Key,继续使用 AI,不消耗平台试用额度。目前支持智谱开放平台和 Coding Plan 的 GLM-5.3,两种服务的密钥和权益不能混用,费用或用量由你的智谱账户承担。密钥由文脉服务器加密保存,不是浏览器直连。
账户名为 2–24 位中文、字母、数字、下划线或短横线;密码为 10–128 个字符。目前不能找回密码,同一网络的试用账户数量也有限制。
注册或查看账户
导入失败怎么办
知乎、公众号等网页可能无法直接提取。可以复制正文,或手动安装 Chrome / Edge 版“网页捕手”。扩展会尝试展开、滚动你有权访问的页面,再读取正文;不能绕过付费或登录限制。导入后请检查是否完整。扫描版 PDF 暂不支持文字识别。
网页捕手安装方法
请勿导入机密或敏感材料
请确保你有权处理文章。使用 AI 时,相关原文、问题和必要的对话会经文脉服务器发送给模型服务,详见数据与隐私。
重要结论需要核实
AI 可能误解原文、漏掉条件或给出错误信息。重要结论请对照原文和可靠来源核实,不要只凭 AI 回答作出医疗、法律、财务等决定。
保存和删除资料
阅读资料默认保存在当前浏览器。你可以在设置中选择文章并同意开启云端档案;进入档案后的新文章和改动会自动上传。换设备需登录同一账户,再打开已有云端档案;之后点“同步并接续”获取远端更新,不会自动覆盖正在输入的内容。这不是端到端加密服务。旧资料不会自动上传,跨篇统计和显示设置暂不跨设备同步。容量与删除范围见数据与隐私说明。
清除站点数据、设备故障或未完成的同步可能导致本机内容丢失。设置中的“个人档案”可备份和恢复文章、问答、草稿、阅读进度及选定的伴读方法;选择全部文章可包含跨篇记录。备份文件未加密,重要内容请另存并妥善保管。
退出登录、清空本地数据不会删除服务器上的账户和用量记录。目前没有自助注销入口,也未设置这些记录的自动删除期限。
其他限制
试用期间功能可能调整,服务可能中断或停止。目前没有正式客服或可用性保证。
Usage guide
Updated October 4, 2026 · Applies to the online trial
What you can use right away
- Samples: no account needed. Article analysis and example answers are prewritten, not generated live.
- Importing and reading: no account needed. Supports pasted text, TXT, Markdown, text-based PDFs, and some public URLs.
- Online AI: requires sign-in. Use the platform trial without model setup, or choose your own Zhipu API key. In both cases, Wenmai's server calls the model.
Trial length and allowance
New accounts currently receive 24 platform AI requests over seven days; check your account page for the actual allowance. One action may involve several requests, so this does not mean 24 questions. After expiry or exhaustion, platform trial requests stop, but saved articles and answers remain readable. There is currently no paid renewal option.
Save and enable your own Zhipu API key in Settings → Model service to keep using AI without platform trial calls. GLM-5.3 on Zhipu Open Platform and Coding Plan is supported. Their keys and allowances are separate; your Zhipu account covers fees or usage. Wenmai stores the key encrypted on its server; the browser does not call the provider directly.
Usernames accept 2–24 letters (including Chinese characters), numbers, underscores, or hyphens. Passwords must be 10–128 characters. Password recovery is not available, and trial accounts per network are limited.
Register or view your account
If importing fails
Sites such as Zhihu and WeChat may block direct extraction. Paste the article text, or manually install Web Catcher for Chrome or Edge. The extension tries to expand and scroll pages you are allowed to access before reading their text. It cannot bypass paywalls or sign-in restrictions. Check the imported article for missing content. Scanned PDFs do not yet support text recognition.
How to install Web Catcher
Do not import confidential or sensitive material
Make sure you have permission to process the article. Using AI sends relevant source text, questions, and necessary conversation through Wenmai's server to the model service. See Data & privacy.
Check important conclusions
AI can misread the source, miss conditions, or give incorrect information. Check important conclusions against the original and reliable sources. Do not make medical, legal, financial, or similar decisions based only on AI answers.
Saving and deleting data
Reading data stays in this browser by default. Select articles and consent in Settings to enable a cloud archive. New articles and changes within that archive upload automatically. On another device, sign in to the same account and open the existing archive. Choose “Sync and continue” for later remote updates; they do not automatically replace what you are typing. Cloud storage is not end-to-end encrypted. Existing device data does not upload automatically; cross-article statistics and display settings do not sync yet. See Data & privacy for capacity and deletion details.
Clearing site data, device failure or incomplete sync may cause device data loss. “Personal archive” in Settings can back up and restore articles, Q&A, drafts, progress and selected companion methods; selecting all articles allows cross-article records too. Backup files are unencrypted. Export important content and keep it safe.
Signing out or clearing local data does not delete server-side account and usage records. There is no self-service account deletion or automatic retention deadline for these records.
Other limits
Features may change during the trial. The service may be interrupted or discontinued. There is currently no formal customer support or availability guarantee.