First @Human Community Meetup — Summary

First @Human Community Meetup — Summary

Hello everyone,

Hope this finds you well.

As I mentioned in the pre-meetup thread, I’m posting this summary of the first @Human Community Meetup so anyone who couldn’t attend has access to what was discussed.

Thanks again to @nmccarthy and @jfortun for hosting, for being transparent about the direction of the project, and for organising the meeting. Thanks also to the other community members who took the time to join.

A request: if anyone who attended (especially @nmccarthy and @jfortun) spots anything I’ve mis-stated or missed, please correct it in this thread.


Topics covered

  1. The @Human project (and Corteza’s transition into it)
  2. Future of Corteza (LTS)
  3. New features in @Human — Agentic, TAQ, Connections, Forms, Discovery
  4. Security considerations
  5. Community interaction going forward
  6. AIRE — current status

1. The @Human project

@Human is a fork and rebrand of Corteza. The team explained that this change was prompted by brand ownership considerations (related to Planet Crust joining a foundation), combined with the opportunity to introduce new AI-related capabilities under a new identity.

Key points:

  • @Human is backwards-compatible with Corteza. Existing functionality is preserved.
  • The main migration work for existing users is the frontend. @Human uses Vue 3 (Vue 2 has been deprecated, and Bootstrap is no longer used). Any custom page blocks or frontend customisations will need to be updated to Vue 3.
  • @Human is currently in early access — still actively being developed, but available to try.

Resources:


2. Future of Corteza

Corteza moves into 18 months of Long-Term Support. During this period:

  • Focus is on bug fixes, security, and performance
  • No new features will be added to mainline Corteza
  • New-feature pull requests for Corteza will not be prioritised — any new work should target @Human
  • Bug fix and security PRs are still welcome

If you’re planning to introduce new functionality, plan around @Human rather than Corteza.


3. New features in @Human

All of the new agentic / AI functionality can be turned off. If you don’t need it, @Human behaves essentially like the Corteza you already use.

Agent Builder

A configurable AI agent system. You can define:

  • Provider (Mistral is the current default; Claude is in testing but usable)
  • Model
  • Token limits
  • Scripts the agent follows
  • Workflows the agent can use as tools
  • The information / data the agent has access to
  • Guardrails
  • Security measures (TCL / act compliance)

Agents will be available with role-based rules.

TAQ — Trigger Action Query

A new, simpler workflow primitive:

  • Designed to be less error-prone — most options selectable via dropdowns
  • Intended for simpler workflows, particularly ones agents will use as tools
  • Documentation focus is more accessible than the existing workflow docs
  • Existing workflows continue to work — TAQ is additive, not a replacement
  • Behind the scenes, TAQs and the existing workflow engine share the same internal logic, so the two interoperate naturally
  • New trigger types are planned (e.g., Git events, Google Docs changes)

Connections / Integrations

A new page for connecting to external services (Git, Google services to start), which can then be used as triggers or data sources in automations. Connections start working immediately after creation — no need to restart the @Human container.

Forms

A form-builder feature for collecting external input. You’ll be able to share a link to a form (similar to a Google Form) and have responses captured directly into a module — for example, straight into your CRM.

New Discovery

The team acknowledged that the current Discovery app is the weakest part of the platform and has received a significant overhaul to improve its functionality and integrate with the AI features. They’re hoping to demonstrate it in the near future.


4. Security considerations

With the new external integrations (Google services, AI providers) and agent execution, security has been a focus area for the team:

  • Permission models for which users can use which connections
  • Avoiding leakage of sensitive information through AI providers
  • Guardrails and policy controls within the agent builder

During the meeting @Toocky / Alex raised useful points about edge cases that still need attention, and Niall acknowledged these and asked Alex to share resources the engineering team could look into. The team is treating these issues seriously and is aware that more points will surface as the integrations mature.


5. Community interaction

The team openly acknowledged that engagement with the community has been reduced over the past year, attributing this to the team being stretched while building @Human.

Going forward:

  • No public roadmaps. The team prefers flexibility to adapt to a fast-moving AI landscape rather than committing to specific milestones they may need to walk back.
  • Monthly community meetups are planned. The next one is expected in the next 2–3 weeks and will focus on @Human again.
  • Direct project-management visibility. The team is working on giving community members access to their @Human-based project-management instance, where current issues and planned work will be visible, and where community members can raise issues directly.
  • Community contributions. Niall expressed interest in getting the community more involved in development. He noted this is harder than it sounds in the current phase because most of the active work is on foundational / internal systems rather than user-facing features that community contributors are typically more comfortable with.

6. AIRE

AIRE is being discontinued. Its functionalities will be migrated to @Human over time. Some are already available through the new agentic features.


Topics I wasn’t able to ask in the time available

I had several questions I didn’t get to in the meeting — mostly more technical or customer-experience-focused topics. Posting them here in the hope they can be addressed in this thread, at the next meetup, or in a follow-up. If anyone else has questions to add, please drop them in below.

  1. Support model going forward. What level of responsiveness should customers realistically expect by issue severity? Is the support model changing with the move to @Human, and is there a published SLA / escalation path?

  2. Environment migration (long-standing pain point). What’s the officially recommended pattern for moving namespaces and modules between dev / staging / production without breaking references due to ID conflicts? Is the existing provisioning feature the intended solution, and is there documentation on best practices?

  3. Performance and scaling. What are the known performance ceilings — record counts, concurrent users, workflow throughput — at which teams typically need to start tuning? Any recommended patterns for modules with 50k+ records?

  4. Customisation patterns. What’s the recommended pattern for heavy customisation in @Human, and at what point do teams typically need to go beyond plugins?

  5. Architecture from day one. If a team were starting on @Human today for a serious internal business platform, what architecture and practices would you recommend from day one to avoid pain later?

  6. Fit assessment. What types of projects are not a good fit for @Human / Corteza, where a different platform would honestly be a better choice?


Thanks again to everyone who joined and to the team for hosting. Looking forward to the next meetup.

Kind Regards,
João Figueiredo

Wow. Many thanks @joaoFigueiredo. That’s an excellent summary and much appreciated.

@Toocky did share some links:

These were the n8n nodes i mentioned.
n8n/packages/nodes-base/nodes at master · n8n-io/n8n · GitHub

My initial response is that we’ll take a look, but that @tjerman did quite bit of research and I’m not sure whether the N8N stuff would be a total change in direction for him (waiting for feedback).

However, one of the hardest things for us to support will be the Connection ecosystem since there are so many elements outside of our control and so many potential corner cases. I think we’ll probably bring this up for more detailed discussion soon.

For our internal backlog, could I ask you all to sign up here please:
https://www.planetcrust.com/human-community

Please say hi in the “Welcome and Introductions” channel by mentioning “@Community” in the chat. I’ll add you into the Human Teams Namespace once I see this. The current Corteza platform backlog looks massive (2500 feature requests), but most of that is because the Connections work has so many bits and pieces.

@joaoFigueiredo for support SLA, I’ll discuss more once you’re in the @Human Community instance. I’ll add the document in there for everyone.

@joaoFigueiredo will reply on the remaining/unanswered topics tomorrow. It may be worth dedicating an entire session to this and asking @gesz and @kornel_david to come along. Bunch of devops related things there.

Hi @nmccarthy,
Thanks a lot for your response. I’ve signed up at the platform as requested.
A dedicated session with @gesz and @kornel_david sounds great — really appreciate you organising that.
Happy to make myself available whenever works for the team.

Whether the session is dedicated or open to the wider community, I’m happy to write up a summary like the this one afterwards so the information is available to anyone who couldn’t attend.
Looking forward to it.

Kind regards,
João Figueiredo