仪表盘

找到失败的请求,验证修复。

所有设备汇入一张实时表格。暂停请求、返回伪造响应,或重放它。

A three-pane dashboard: device rail, live request feed with status codes and detail pane. One row is lifted out on a paused breakpoint badge being edited, another returns a mocked 500, and two pinned rows connect to a side-by-side diff card whose changed lines carry red minus and green plus markers. POST GET 200 paused GET 200 GET MOCK 500 PUT 201 { } - +

流量工作台

Find it. Fix it.

筛选到你关心的那个请求,然后修改或重放它。

实时消防水龙带

每个请求落地瞬间即流入一张可排序表格。没有刷新按钮。

什么都能模拟

用你自己的状态码、标头和正文应答匹配的请求——或直接丢弃。

在线路上写脚本

JS 钩子实时改写标头与正文,配有 dry-run 测试器,坏脚本永远不会拖垮捕获。

断点

暂停匹配的流量,编辑原始 HTTP,继续或重发。

好坏对比,一目了然

固定条目,并排对比原始请求与响应。

导出、定时、解码

导出 HAR、定时快照、解码 protobuf 与 gRPC。

How it works

Device to dashboard, one hop

No upload step and no poll loop — capture writes a row, Realtime fans it out, and control flows back down the same wire.

  1. 01

    Every source writes the same row

    iPhone tunnel, Android VpnService, the MITM proxy, Chrome over CDP, farm phones — each captured request becomes one identical entry row the moment it completes. One shape, every platform, no importer.

  2. 02

    Realtime pushes it to the feed

    The dashboard holds a WebSocket subscription to your workspace's broadcast channel; a committed entry lands in the live table without a refresh or a poll loop.

  3. 03

    Control rides the same wire back

    Hold a matching request at a breakpoint, edit the raw HTTP, continue or abort — or resend a past request. The capturing device answers over its own Realtime channel in seconds.

  4. 04

    Settings save once, apply everywhere

    SSL lists, block and mock rules, and rewrite scripts push live to every affected device — globally, per service group, or per device. No redeploys, no app updates.

The surface

More than a feed

The feed is the front door; the same workspace carries explorers, boards and device controls for the whole stack.

Capture sources
iOS · Android · MITM proxy · Chrome (CDP) · device farm
Feed transport
通过持久 WebSocket 实时推送——条目在提交的那一刻即出现,绝不依赖轮询定时器
持续吞吐量
每天有 205,000 条捕获请求——约 8.81 GiB 流量——流经实时信息流(生产集群,2026 年 7 月实测)
Workspace model
Per-user tabs with saved filters; Devices, Services and Domains panels scope the feed
Interception
breakpoints · mocks · block rules · JS rewrite scripts
API explorers
MCP · REST · WS — self-documenting in-app consoles
Team boards
todo · roadmap · farm · tester
Device control
Live screen share with saved recordings, remote VPN toggles, renames and health — per device
Access control
Role-based permissions and workspace scoping, enforced by the database itself (row-level security)

FAQ

Dashboard questions, answered

Do I need to install anything to use the dashboard?

No. It's a web app — sign in and the live feed starts the moment any paired device or proxy client captures traffic. Capture happens on the devices; the dashboard is the control room.

How fast do entries show up?

Entries stream over a persistent WebSocket as rows commit — the feed updates in real time, with no refresh button and no polling interval to tune.

Can my whole team watch the same traffic?

Yes. The feed is workspace-scoped, so every member sees the same devices side by side, while tabs and filters stay per-user and role-based access control decides who may view or change what.

Can AI agents work with the same data?

Yes — every dashboard read and write is mirrored by the MCP server, the REST surface and the WS channels, so an agent can search entries, flip settings or continue a breakpoint with the same permissions you have.

See your traffic live

看着请求从每台已连接设备流入——然后聚焦、筛选、断点、模拟。

Ask your mate