Digital Meld

Veritas Kanban, our new, open source agentic AI orchestrator went live today. Visualize and plan your agents flows! We have nine new features planned for v1.1, and would love for your feedback or assistance. All aboard the Clawdbot/Moltbot hype train!🚂 github.com/BradGroux/veri…

Thank you to @github for having my back. GitHub Copilot Pro+ is THE best value on the market. github.com/features/copil… I have been able to review, test and comment on eight (and counting) Microsoft related PRs for OpenClaw tonight thanks to @Meta_Ashley!

Veritas Kanban just hit v4.0 and 600+ stars on GitHub. When I started building VK, it was a side project to scratch my own itch, for getting to know OpenClaw. I needed a task board that actually worked with AI coding agents instead of fighting them. Markdown files, local-first, no cloud dependency. v4.0 adds the governance layer I've been wanting to build for weeks: • Agent Policy Engine: configurable guard rails for what agents can and can't do • Decision Audit Trail: every agent decision logged with assumptions and outcomes • Behavioral Drift Detection: catch when agents start going off the rails • Output Evaluation Scoring: quality metrics on what agents produce • Customizable Dashboard with drag-and-drop widget grid • Prompt Registry with full version control • 33+ MCP tools for Claude, Codex, Copilot, and other MCP-compatible agents Plus, a self-hosting guide, 11 new operational SOPs, and a complete API reference. Big thanks to the community members who contributed work that made it into this release. 623 stars, 60 forks, and growing. If you're running AI agents on real projects and want guardrails that actually work, give it a look! github.com/BradGroux/veri…

If you care about the future of OpenClaw, be sure to follow Dave - he's on the board with @steipete and they are working together to protect the future of the project. Check out his conversation about the Foundation with @jason and @TWiStartups - youtube.com/watch?v=KEHMoI…

The last 48 hours have been a whirlwind. What started as a simple effort to improve the Microsoft Teams plugin for OpenClaw has turned into something bigger, thanks to the passion and support from people inside Microsoft, my fellow maintainers, and all of you following along with us! At this point, I’ve been connected with well over 100 Microsofties through this process (and spoken directly to over 25), and that’s only reinforced my confidence that the right people are paying attention and pushing in the right direction. Folks like @fabianwilliams, @shanselman, and @upster have been especially meaningful and supportive voices in helping move things forward, and I’m grateful to have allies like them in our corner. They're on it, and we're in good hands. Out of respect for the work happening behind the scenes by our allies, I’m just asking for a little patience. Good things are coming! Remember y 'all, 🦞 is the law.

Proud to feature Brad Groux from Digital Meld on the Software Oasis AI Bootcamp stage, sharing how founders can use AI to accelerate digital transformation and growth. @digitalmeld

Question: Are you using Teams with OpenClaw? I'd love to have more data to present to Microsoft, and to my fellow maintainers on the number of folks either using Teams, or who would like to use Teams with OpenClaw.

Well, it looks like I'm now an official maintainer for @OpenClaw. I spent 2 months yelling at my AI assistant through Microsoft Teams before anyone noticed. Back in January, I wrote a how-to guide for connecting #OpenClaw to #MicrosoftTeams because the docs were thin and the setup was painful. Azure Bot Service, Entra ID, Cloudflare tunnels, real enterprise plumbing that most people avoid. Through my interactions with the team, and the article, I caught the attention of @steipete and @onusoz, the team behind OpenClaw and the Teams plugin. I volunteered to come on as a key maintainer for the Teams integration. Today I found and fixed a bug that had been silently dropping every channel message since the last update. Took 3 hours of patching compiled runtime logers because the debug wrappers were swallowing all the diagnostic data. The kind of thing you only find when you're actually running this in production, not just testing it. That's the point. Microsoft Teams is where enterprise lives. If AI agents are going to matter, they need to work where the work already happens, not just in terminals and chat apps built for developers. I'm grateful to Peter and the entire OpenClaw team for building something worth contributing to. Proud to play my small part. If you're running OpenClaw with Teams and the #AzureBotService, or trying to, reach out! I've been through every edge case. And if anyone from the @MicrosoftTeams team is reading this: let's talk. There's a lot we could build together.

Also, call me crazy, but I'm not going to take what is or isn't working for businesses from people that have never actually worked in the business world. I've talked to thousands of small and midmarket business leaders over the past three years, and I can assure you that none of them have ever been asked by a PhD, what is or isn't working for their business. Small and mid-market companies are 50% of the economy, and small businesses make up over 99% of all businesses in the US... Long story short, there's more to it than Silicon Valley, CNBC, and the S&P 500.

Our open source AI Agent orhcestration platform, Veritas Kanban, has reached 500 GitHub stars and has been downloaded over 15,000 times. Thank you so much to everyone who has downloaded, supported, and contributed. I would have never imagined we'd get this far! ⭐️⭐️⭐️⭐️⭐️ It was built with OpenClaw, for OpenClaw, but has grown into a fully agent agnostic platform. Huge news for v4.0 coming soon! github.com/BradGroux/veri…

OpenAI just launched Codex for Open Source, free Pro access, API credits, and security scanning for OSS maintainers. This is how you build an ecosystem! I applied immediately for Veritas Kanban. We're building the orchestration layer that connects AI agents to real work. Task management, telemetry, multi-agent coordination... The stuff that makes agentic AI actually useful in production. 488 ⭐ and climbing! If you're building with AI agents, this is the missing piece. Big respect to @OpenAI for backing the people who build in the open, and thanks to @steipete for all he has and continues to do for the Open Source movement. 🫡 github.com/BradGroux/veri…

Most businesses don’t have an AI problem, they have a workflow and structure problem. In my Software Oasis Virtual AI Advantage Bootcamp session and Q&A panel, I shared a practical framework for getting real results from AI agents: • Build structure before speed • Use PRDs and SOPs to define expectations • Treat markdown as your operational memory • Start with personal workflows, then scale to teams If you’re trying to move from AI experimentation to operational impact, this talk is for you. #AIagents #AgenticAI #OpenClaw #BusinessTransformation

RT @BradGroux: Something shifted in the past few months. I've been building software for years. I've seen waves come and go. But what happe…

RT @BradGroux: Holy smokes, Veritas Kanan has 400 stars on GitHub. I'd had never thought that my first ever open-source project would do so…

I've been obsessed with documentation for over two decades. PRDs (Project Requirements Documents) before writing code. SOPs (Standard Operating Procedures) before shipping features. My teams used to roll their eyes. "Brad, we don't need another doc. Just let us build." So I kept quiet about it. Wrote them anyway. Filed them where no one would complain. Then AI agents happened. Last month I pointed an AI agent at a PRD I wrote years ago. It broke the spec into 14 tasks, estimated complexity for each one, identified three edge cases I missed, and started building. No prompt engineering. No clever tricks. Just a clear document it could actually read. I tried the same thing with our deployment SOP. The agent runs it at 3am now. Every step. Every verification. No "oh I forgot that part." No drift. The results so far: → 80% less time re-explaining decisions → Agents that ship what I actually meant → Docs that compound instead of collecting dust → Cost savings on token usage Here's the part nobody talks about: The best AI users aren't the ones with the cleverest prompts. They're the ones who've been writing clear documentation for years. All those PRDs and SOPs I hoarded? They're not overhead anymore. They're executable infrastructure. Turns out the boring stuff was the leverage all along. More details found in the comments.

Episode 0034 of the Start Small, Think Big podcast is now live. Vibe Coding vs Real Systems: The Truth About Shipping AI and AI Agents in Production This was 24-hours after learning about Clawdbot / MoltBot / OpenClaw! youtube.com/watch?v=YXzrML…

10x productivity isn't about better AI. It's about better systems. Just gave a lightning talk at Houston AI Club on how @DigitalMeld shipped 909 tasks, 13 releases, and achieved 100% agent success rate in 11 days. Here's what actually worked with OpenClaw, Claude Code, and Codex. Most teams use agents wrong. They throw vague prompts at Claude and hope. That doesn't scale. What scales: PRDs. SOPs. Guardrails. Clear contracts that agents can actually follow. Agents aren't a replacement for structure, they're a multiplier of it. That's where Veritas Kanban comes in. Open source. Built by agents, for agents. 909 tasks in 11 days. Not because the AI got smarter. Because the system did. PRDs > prompts. SOP > shortcuts. Every time. github.com/BradGroux/veri…

OpenClaw has transformed our business here at @DigitalMeld, that isn't hyperbole, it is real. The best part is, our improved capacity is passed right on to our customers at no additional cost - they're being rewarded for being with us before this wave of innovation.

Six years ago, I started building with AI because I believed it could help people punch above their weight. This week, three things happened on the same day: • An OpenAI engineer publicly said he "finally feels the existential threat." • An Anthropic safety researcher quit, saying "the world is in peril." • And a startup founder wrote that he now describes what he wants in plain English, walks away for four hours, and comes back to finished work, better than he'd have done himself. These aren't pundits. These are the people building the thing. I run an AI automation company. I use these tools every single day. And I'll be honest, the gap between what I tell people at dinner and what's actually happening has gotten uncomfortable. Here's what I know: the companies that figure out how to work WITH this wave will thrive. The ones waiting for it to blow over are already behind. The question isn't whether AI will reshape your industry. It's whether you'll be the one reshaping it, or the one being reshaped. If you're a leader who hasn't sat down with your team to talk about this yet, that conversation is overdue.

Two weeks ago I built a kanban board for my OpenClaw AI agents. Tonight it became a full orchestration platform. Veritas Kanban v3.0 ships a complete workflow engine: → YAML pipelines that chain agent work across steps → Loop, gate, and parallel orchestration → Human approval gates (agents can't ship without you) → Role-based tool policies (reviewers can't edit code they're reviewing) → Real-time dashboard monitoring every run The Claude and Codex agents that built it? They reviewed each other's work. Cross-model. Every phase scored 10/10/10/10 across code quality, security, performance, and architecture. 22,500 lines. 8 issues. 4 phases. One session. This isn't a task board anymore. It's how AI teams actually ship software. Open source → github.com/BradGroux/veri… #AI #OpenSource #AgentOrchestration #BuildInPublic

I shipped 8 releases in 7 days. Well... I didn't do it alone. Meet VERITAS : Virtual Enterprise Resource for Intelligence, Tasks, Analysis & Strategy. My AI agent who co-built Veritas Kanban v2.0 with me. Not a copilot. A teammate. VERITAS wrote the agent registry. Built the dashboard widgets. Tracked token usage, and cost per task spend. Ran security audits. Filed GitHub issues. Debugged production at 2 AM. The result? A full multi-agent orchestration platform: • 87K lines of TypeScript • 71 issues closed • 1,260 tests passing • 288 ⭐ in a week When people ask "can AI really build software?" - this is my answer. Not a demo. A shipped product. github.com/BradGroux/veri… #OpenClaw #Claude #Codex #AgenticAI #AIagents