WorkMesh

English | 简体中文

Make every capability visible. Help every person affected by the AI transition find the next meaningful opportunity faster.

WorkMesh, also referred to as the Future Work Platform, is an open-source prototype for future employment and workforce collaboration.

It explores how people, skills, jobs, enterprise demand, service fulfillment, governance operations, and AI-assisted decision support can be connected through one shared data loop. The goal is not to build another job board. The goal is to explore a more efficient, transparent, and explainable coordination network that offers more dignified support to people navigating unemployment, career transitions, flexible work, and continuous learning.

A Note from the Initiator

I am an ordinary humanities graduate, not a professional software engineer. Most of the code in this repository was completed with the help of AI. I was surprised by what AI could do, and I also felt the shared power behind Hello World. That led me to explore a skill-sharing system built around the idea of "open-source skills." As more people face unemployment or career disruption during the AI transition, perhaps a system like this can become useful.

For the system to work in the real world, it would need backing from a government body or a large organization with strong public credibility. It aims to respond to several challenges: supporting people who lose their jobs, reducing resource imbalances between enterprises of different sizes, and improving oversight of capital flows.

This is as far as I can take the system with my current abilities. If you would like to contribute to the future, you can improve it through technology or share it so more people can prepare earlier.

Why This Exists

Traditional employment systems often split the same person into disconnected records: a resume on a recruitment website, training history on another platform, job-seeking status scattered across different systems, and delayed statistical reports on the governance side.

WorkMesh explores a different model:

1. A person registers once and receives a unique identity reference through a trusted government body or large organization, allowing a durable personal profile to grow over time. Future exploration may include Web3 technologies to make identity more trustworthy, unique, and resistant to tampering.
2. An enterprise registers once and receives a unique identity reference through a trusted government body or large organization, allowing enterprise information to accumulate over time.
3. A government body or trusted organization can use monitoring, governance, and allocation mechanisms to understand employment operations and encourage a more balanced flow of resources.
4. The platform can turn skill gaps into learning paths and next actions.
5. AI assistants can retrieve platform knowledge and invoke role-aware business capabilities.
6. Governance teams can understand employment operations through shared metrics, risks, tasks, and audit trails.

The intended result is a living employment collaboration network: when one profile changes, the rest of the system can respond.

Platform Overview

This repository contains a locally runnable multi-portal system:

• A Personal Web Portal for workers, job seekers, and flexible talent.
• An Enterprise Portal for recruitment, supply-demand collaboration, fulfillment, assets, and finance.
• A Governance Portal for employment analysis, risk monitoring, and closed-loop resolution.
• A Mobile Registration Portal that creates an account and a uniquely linked personal profile in one flow.
• A Shared API for authentication, matching, RAG retrieval, Agent Skills, and cross-portal data access.

Portal Map

| Service | Port | Entry URL | Primary Users | AI Surface | | --- | ---: | --- | --- | --- | | Enterprise Portal | 3000 | http://localhost:3000/enterprise/dashboard | Recruiters, operators, enterprise admins | Enterprise AI Assistant | | Personal Web Portal | 5173 | http://127.0.0.1:5173/#/ai-assistant | Workers, job seekers, flexible talent | Opens in the Personal AI Workspace | | Governance Portal | 8000 | http://127.0.0.1:8000/#/en/cockpit | Employment service and governance teams | Bodao AI | | Mobile Registration Portal | 5174 | http://127.0.0.1:5174/#/register | New users | Hands off to the Personal AI Workspace | | Shared API and Swagger | 3001 | http://localhost:3001/api/docs | Developers and integration partners | RAG and Agent Skills endpoints | | PostgreSQL | 5432 | localhost:5432 | Shared infrastructure | Stores profiles, matching, RAG, and invocation audit records |

In the table above, localhost and 127.0.0.1 refer to the visitor's own computer. Publishing the repository does not automatically create a public website. After downloading the repository and running ./deploy.sh or ./start-local-demo.sh, users can access the system locally through these URLs.

When deploying WorkMesh to a cloud server, replace localhost with the server IP address or your own domain, for example http://203.0.113.10:3000/enterprise/dashboard. For production deployments, configure HTTPS and reverse proxy routing through Nginx, Caddy, or a cloud gateway. Separate subdomains such as enterprise.example.com, personal.example.com, governance.example.com, and api.example.com are recommended.

Architecture

flowchart LR
    Person["Personal Web Portal<br/>React + Vite<br/>5173"] --> API["Shared API<br/>NestJS + Prisma<br/>3001"]
    Register["Mobile Registration<br/>React + Vite<br/>5174"] --> API
    Enterprise["Enterprise Portal<br/>Next.js<br/>3000"] --> API
    Governance["Governance Portal<br/>HTML + CSS + JS<br/>8000"] --> API
    API --> DB[("PostgreSQL<br/>5432")]
    API --> Match["Skill and Job<br/>Matching Engine"]
    API --> RAG["RAG Knowledge<br/>Documents"]
    API --> Skills["Agent Skills<br/>Router"]
    Skills --> Audit["Agent Invocation<br/>Audit Trail"]
    Match --> DB
    RAG --> DB
    Audit --> DB

Design Principles

• One profile, many opportunities. Registration creates a uniquely linked User and PersonalProfile.
• Structured capability first. Standardized skill tags provide the basis for explainable matching.
• Cross-portal collaboration by default. Personal profiles, enterprise jobs, and governance metrics share one database.
• AI is more than a chat window. Assistants can retrieve knowledge and route requests to business Skills.
• Traceability matters. Agent calls and governance workflows can leave audit records.
• Prototype honesty. The repository presents the product direction while making production hardening work explicit.

Core Product Concept

The following mechanisms describe the long-term direction of WorkMesh. The current repository already implements some of them. Features that are not yet complete are marked clearly later in this document.

Personal Profiles and Open-Source Skills

• Skills become recognizable tags. Whether someone works freelance or full time, they can continue adding their skills. When demand appears, the system can identify more suitable people.
• Learning continues when skills are missing. Users can study a target skill or publish a learning request so someone with that skill can teach them. After assessment, the learner can also receive the corresponding skill tag.
• Everyone can share skills. People can open-source what they know, teach others, and receive compensation. Knowledge and experience can keep moving between people. From Hello World to Hello Skill.
• AI works through platform capabilities. The assistant can call platform APIs instead of behaving like an isolated chatbot. Recommendations can take a person's current state into account, such as busy, available, or temporarily not working.

Example: if someone has an "electrical circuit repair" skill, and an enterprise or individual publishes a matching request, a future version of the system could combine skill tags with a nearby 30km distance search.

Enterprise Profiles and Supply-Demand Collaboration

• Enterprises become recognizable tags. Business scope, qualifications, and capabilities can become detailed tags. Professional and trusted tags should take effect only after governance-side review.
• Enterprise scale is linked to fulfillment capacity. The amount of orders an enterprise can accept through the platform can be connected to factors such as office space, staffing, and fulfillment capability. To increase its revenue limit, an enterprise should add the corresponding talent and production capacity.
• Skill-based recruitment connects to the Personal Portal. When an enterprise publishes a job, the platform can match suitable talent and help fill capability gaps quickly. Future versions may introduce recruitment deposits or similar trust mechanisms to reduce false job listings.
• AI connects supply and demand. Enterprises can publish collaboration offers and demand requests. AI can assist matching based on qualification tags and enterprise tags.

Example: if an enterprise has a "lawn mower manufacturing" tag, and another enterprise or individual needs a product, a large order, or production support, a future version of the system could prioritize qualified enterprises in the same city and encourage regional collaboration.

Governance and Public Services

• Trusted identity. Provide unique identity references for people and enterprises to reduce duplicate entities and fa