ProToDo
Speak a paragraph and AI turns it into a three-level task board — clients see live progress from a link, no account needed.
In short
ProToDo is a project-collaboration and progress-alignment platform for software agencies and their clients. Admins dictate or paste requirements and AI organizes them into a three-level board — phase, task, subtask — with status rolled up automatically. What sets it apart: clients need no account at all; a single login-free link lets them check progress, comment, and confirm, and a one-click PDF summary closes out the project.
What the agent takes off people's plates
Core capabilities
Dictate, transcribe, auto-organize
The browser-native Web Speech API transcribes speech in real time; you can edit the text before AI breaks it into structured tasks. Pasting meeting notes or quickly stacking loose to-dos works the same way.
Phase / task / subtask board
Five task states, assignees, due dates, notes, and drag-to-reorder. Status rolls up automatically from subtask to task to phase to project, so the progress bar and status badge never disagree.
Login-free client share page
Each project gets a read-only `/s/<slug>` page. Clients check progress without signing up, installing an app, or joining a group chat — and they can comment and confirm right there. Links can carry an expiry or be switched off anytime.
One-click project summary PDF
The full lifecycle — phases, tasks, timeline, client feedback — is compiled into a summary document and exported as PDF via browser print, ready to serve as the acceptance and close-out deliverable. No more hand-written wrap-up reports.
Project file sharing (Tencent COS)
Every project has its own file area; upload, download, and delete go through a server relay with signed URLs, and clients can pull deliverables straight from the share page. Without COS configured the file tab degrades gracefully instead of breaking.
Self-hosted stack, data stays yours
Next.js 15 App Router + React 19 + strict TypeScript, with data in a local SQLite file (better-sqlite3) and sessions on httpOnly cookies with scrypt-hashed passwords. `output: standalone` emits a self-contained server, so Docker deployment is a single command.
Side by side
ProToDo vs. general project-management SaaS vs. group chat + Excel
| Dimension | ProToDo | General PM SaaS (Jira / Teambition) | Group chat + Excel |
|---|---|---|---|
| Barrier for the client | None: no signup, no app — open the link, read, comment, confirm | Requires an account in your workspace, usually consuming a paid seat | No barrier and no structure — progress is asked for message by message |
| Progress visibility | Subtask → task → phase → project rolls up automatically; client and team read the same live state | Capable, but views and permissions must be configured first; clients usually receive only exported reports | The sheet is stale the moment it is sent; mismatched versions must be reconciled by hand |
| Data ownership | Self-hosted: data lives in a SQLite file on your own server and can be taken with you at any time | Data sits in the vendor cloud; export format and retention are the vendor’s call | Scattered across personal chat histories and local files — it walks out with whoever leaves |
| Deliverable trail | One click produces a summary PDF covering the full lifecycle: phases, tasks, client feedback | Reports and charts exist, but the close-out document is still written by hand | Almost nothing persists; retros mean digging through chat logs and multiple Excel versions |
Quantified before / after
Measured impact
Verifiable facts
Numbers and sources
Work is organized as phase → task → subtask, tasks carry five states, and status rolls up from task to phase to project so progress and status badges always agree.
The login-free client page lives at `/s/<slug>`: read-only progress with comments and confirmations, and the link can carry an expiry or be switched off.
AI organization defaults to a built-in heuristic requiring no API key; once an OpenAI-compatible Base URL and key are entered under Settings → AI Models, the real provider takes priority and failures fall back to the heuristic.
Dictation uses the browser-native Web Speech API for real-time transcription; the transcript can be edited before AI turns it into tasks.
The stack is Next.js 15 (App Router) + React 19 + strict TypeScript, with persistence in SQLite (better-sqlite3) and auth via httpOnly-cookie sessions with scrypt-hashed passwords.
Production builds use `output: standalone`, emitting a self-contained server that bundles its dependencies (including the better-sqlite3 native module); in Docker the SQLite file defaults to the persistent volume path `/data/protodo.db`.
Project file sharing runs on Tencent Cloud COS with a server relay and signed URLs, configured under Settings → Object Storage; when unconfigured, the file feature degrades gracefully.
Delivery and pricing
Pricing
Included
- · Full self-hosted deployment license, perpetual use, with no cap on projects, clients, or team members
- · Docker standalone deployment, database seeding, and guidance for reverse proxy and HTTPS setup
- · Training for admins and team members: dictation-based organization, board management, share-link and expiry policy, summary PDF export
- · One year of technical support and version updates, including help with brand theming (theme, accent color, corner radius, fonts)
- · Bilingual Chinese/English interface out of the box
Provided by the client
- · Server and persistent disk (the SQLite database must live on a persistent volume)
- · Domain name and HTTPS certificate
- · Tencent Cloud COS account plus its storage and traffic fees (required for project file sharing)
- · For real-model organization, your own OpenAI-compatible Base URL, API key, and token costs (skip it and the built-in heuristic is used)
- · Custom development beyond the standard feature set and integrations with third-party systems
Add-ons
- · Renewal of technical support and version updates from year two onward
- · Custom development billed per person-day (new modules, approval flows, WeCom / DingTalk notifications, etc.)
- · Data-layer migration from SQLite to managed PostgreSQL, enabling multi-instance or serverless deployment
- · Private-model integration and organization-prompt tuning against your self-hosted OpenAI-compatible gateway
Scan for the full solution document
Scan with WeChat to reach the solution lead and request the complete documentation and deployment checklist.
Where it fits
When this approach does not apply
- Requires deployment with a persistent disk. ProToDo is a long-running Node service backed by a file-based database (SQLite), suited to a VPS, self-hosting, or Docker with a persistent volume. It is not suitable for stateless serverless platforms unless the database is migrated to a managed one.
- AI organization defaults to a built-in heuristic and works with zero keys. Connecting a real model means supplying your own OpenAI-compatible Base URL and API key in settings; failed calls fall back to the heuristic, so organization quality tracks whichever model you connect.
- Speech transcription relies on the browser-native Web Speech API and is bound by browser implementation and network conditions. Browsers without the API fall back to pasting or typing, and accuracy drops in noisy environments.
- Project file sharing needs your own Tencent Cloud COS. Until an admin enters SecretId / SecretKey / Bucket / Region in settings, the file feature degrades gracefully: everything else keeps working, but deliverables cannot be uploaded or downloaded.
Further reading
Related applications
DataWeaver
Five specialized agents cut cross-database natural-language queries from 2–3 days to 10 seconds.
Meeting Management — Private-Deployed Platform
A ¥19,800 perpetually licensed, self-hosted meeting platform — 8 modules, any BCP 47 language, 4-channel international payment, AI site builder + Copilot, dual-QR check-in. v0.6.0 in production.