Skip to content

New — Introducing Yaju Software Factory — Try the tool for free

Explore now
Yaju AS
  • ProductProduct
  • SolutionsSolutions
  • ResourcesResources
  • BlogBlog
  • CompanyCompany
Sign in
Book demo

Platform

  • Agent Orchestration System

    Building agents is easy. Operating should be too.

  • Software Factory

    Turn your backlog into review-ready code

  • Agent Hub

    Browse, run, and share agents

Functionalities

  • AI Governance

    Policy enforced at the moment of action

  • AI Observability

    Observe and trust every agent

  • Token Monitoring

    Make every token count

  • Optimizer

    Same outcomes, lower cost

  • AI Spend Explorer

    Find overspend in two minutes

Product

  • MCP Gateway

  • CLI

  • Pricing

  • Versions

Featured

Choosing a model is an operations decision, not a benchmark decision

Use Cases

  • Cost Control

    Know what agents cost. Prove what they deliver.

  • AI Transformation

    Turn AI adoption into business transformation

Deployment

  • Credential Vault

    Org-level secrets, resolved at runtime

  • Self-hosted

    Run agents in your own environment

By Industry

  • IT & Developers

  • Financial Services

  • Public Sector

  • Engineering

  • Telecommunications

  • Healthcare and Life Sciences

  • Manufacturing

Featured

Running agents on hardware you own

Discover

  • Customer Stories

  • Partners

  • Yaju Labs

    Yaju Agent Systems research lab

For Learners

  • Versions

  • Agent Academy

Featured

Support triage is the best first agent most teams never build

Content

  • Blog

    The latest from Yaju, launches, and insights

Explorations

  • Future(s) of Work

    How will AI change the way we work?

  • Oran Models

    The generation teams run today

Initiatives

  • Scholars Program

    Finding the next generation of agent builders

  • Open Development Community

    Building agent tooling in the open

  • Catalyst Grants

    Backing ambitious work on agents

Featured

The future of work debate has an evidence problem
  • About

  • Careers

  • Newsroom

Yaju AS
Sign in
Book demo

Platform

  • Agent Orchestration System

    Building agents is easy. Operating should be too.

  • Software Factory

    Turn your backlog into review-ready code

  • Agent Hub

    Browse, run, and share agents


Functionalities

  • AI Governance

    Policy enforced at the moment of action

  • AI Observability

    Observe and trust every agent

  • Token Monitoring

    Make every token count

  • Optimizer

    Same outcomes, lower cost

  • AI Spend Explorer

    Find overspend in two minutes


Product

  • MCP Gateway

  • CLI

  • Pricing

  • Versions


Featured

Choosing a model is an operations decision, not a benchmark decision

Use Cases

  • Cost Control

    Know what agents cost. Prove what they deliver.

  • AI Transformation

    Turn AI adoption into business transformation


Deployment

  • Credential Vault

    Org-level secrets, resolved at runtime

  • Self-hosted

    Run agents in your own environment


By Industry

  • IT & Developers

  • Financial Services

  • Public Sector

  • Engineering

  • Telecommunications

  • Healthcare and Life Sciences

  • Manufacturing


Featured

Running agents on hardware you own

Discover

  • Customer Stories

  • Partners

  • Yaju Labs

    Yaju Agent Systems research lab


For Learners

  • Versions

  • Agent Academy


Featured

Support triage is the best first agent most teams never build

Content

  • Blog

    The latest from Yaju, launches, and insights


Explorations

  • Future(s) of Work

    How will AI change the way we work?

  • Oran Models

    The generation teams run today


Initiatives

  • Scholars Program

    Finding the next generation of agent builders

  • Open Development Community

    Building agent tooling in the open

  • Catalyst Grants

    Backing ambitious work on agents


Featured

The future of work debate has an evidence problem
  • About

  • Careers

  • Newsroom

Yaju AS
  • Agent Orchestration System

    Software Factory

    Agent Hub

  • MCP Gateway

    CLI

    Pricing

    Versions

  • AI Governance

    AI Observability

    Token Monitoring

    Optimizer

    AI Spend Explorer

  • Cost Control

    AI Transformation

    Solutions Overview

  • IT & Developers

    Financial Services

    Public Sector

    Engineering

    Telecommunications

    Healthcare and Life Sciences

    Manufacturing

  • Credential Vault

    Self-hosted

    Deployment Options

  • Blog

    Customer Stories

    Partners

    Agent Academy

    Yaju Labs

  • About

    Careers

    Newsroom

  • Legal Center

    Security

    Privacy Policy

    Terms of Use

Platform

  • Agent Orchestration System

  • Software Factory

  • Agent Hub

Product

  • MCP Gateway

  • CLI

  • Pricing

  • Versions

Functionalities

  • AI Governance

  • AI Observability

  • Token Monitoring

  • Optimizer

  • AI Spend Explorer

Solutions

  • Cost Control

  • AI Transformation

  • Solutions Overview

By Industry

  • IT & Developers

  • Financial Services

  • Public Sector

  • Engineering

  • Telecommunications

  • Healthcare and Life Sciences

  • Manufacturing

Deployment

  • Credential Vault

  • Self-hosted

  • Deployment Options

Resources

  • Blog

  • Customer Stories

  • Partners

  • Agent Academy

  • Yaju Labs

Company

  • About

  • Careers

  • Newsroom

Legal

  • Legal Center

  • Security

  • Privacy Policy

  • Terms of Use

Platform

  • Agent Orchestration System

  • Software Factory

  • Agent Hub

Product

  • MCP Gateway

  • CLI

  • Pricing

  • Versions

Functionalities

  • AI Governance

  • AI Observability

  • Token Monitoring

  • Optimizer

  • AI Spend Explorer

Solutions

  • Cost Control

  • AI Transformation

  • Solutions Overview

By Industry

  • IT & Developers

  • Financial Services

  • Public Sector

  • Engineering

  • Telecommunications

  • Healthcare and Life Sciences

  • Manufacturing

Deployment

  • Credential Vault

  • Self-hosted

  • Deployment Options

Resources

  • Blog

  • Customer Stories

  • Partners

  • Agent Academy

  • Yaju Labs

Company

  • About

  • Careers

  • Newsroom

Legal

  • Legal Center

  • Security

  • Privacy Policy

  • Terms of Use

LinkedInInstagramEmail

Yaju AS ©2026

  • English

Research

Community research across regions

A research programme run from one city sees the problems of that city. Most of what we have learned about operating agents in difficult conditions came from members working somewhere with different constraints.

Yaju Team · 4 August 2026

Yaju Labs is a small team in Barcelona. On its own it would produce research shaped by the conditions Barcelona has: reliable infrastructure, a particular regulatory environment, a particular set of assumptions about what a normal deployment looks like.

The Open Development Community is how we avoid that being the whole picture.

What distributed contributors actually surface

Constraints we do not have. Intermittent connectivity, which changes what a retry policy should do. Expensive bandwidth, which changes how much context is reasonable to carry. Regulatory environments with different rules about where data may sit.

Vocabulary and language conditions that a team operating in one language never encounters, particularly code-switching within a single sentence, which most systems handle badly and few evaluate for.

And simple variety of use. An agent doing procurement work in one sector behaves differently from the same agent doing the same nominal task in another, and no amount of internal testing produces that range.

How it works in practice

Members pick up open pieces of work or propose their own, and ship alongside the team rather than around it. Findings are published rather than absorbed.

It is free to take part and it stays free. We do not charge for participation in the Open Agent System Development Initiative, and that is not a launch offer.

What we have changed because of it

The per-language evaluation guidance came from members demonstrating that aggregate quality figures hide exactly the cases that matter, by averaging a well-served majority with an underserved minority.

The emphasis on hybrid retrieval, combining semantic and keyword search, came from contributors working with corpora where embeddings alone handled internal terminology poorly.

Neither would have come from a roadmap.

What we ask

Publish what you find, including the parts that do not support a clean narrative. Research that only reports successes describes a world nobody works in.

Beyond that, nothing. There is no obligation to produce, no reporting requirement and no exclusivity.

Taking part

The Open Development Community page covers how to join. The Scholars Program is a remote, part-time route with a mentor for people who want something more structured, with applications each September.

Next

Yaju Labs, the Open Development Community and the Scholars Program pages cover each route. Questions can go to contact@capconsultor.eu.

Loading...
Community research across regions | Yaju