Blog
Tutorials, news, and release notes.
Showing 1–12 of 21 articles
Marketing & Content6 min read
Beyond the Prompt: How AI Agents Are Becoming Your Content Team
A practical look at how bloggers and developers alike can use AI agents like Claude, connected via MCP, to draft, illustrate, and publish web content end to end — including the real hiccups we hit writing this post.
Engineering4 min read
The Future of AI Agents Isn’t One Model. It’s Orchestration
AI coding agents are powerful, but expecting one model to do every job is inefficient and risky. The next leap is orchestrating multiple models, agents, and tools under clear controls.
Product3 min read
Introducing Archy MCP: Publish Blog Posts and Websites from Any AI Assistant
We shipped @archy-ai/blog-post-mcp — an open MCP server that lets you create, manage, and publish blog posts and entire websites directly from Claude, ChatGPT, VS Code Copilot, Cursor, or any MCP-compatible AI client.
Engineering10 min read
AI Agent Code Review: Workflow, Guardrails and Measurement
A deep dive into structuring AI agent teams for software development — from automated code review and Jira bug filing to self-reflecting agents that improve their own instructions. Based on real production workflows.
Engineering5 min read
Agent Orchestration Playbook: From Task Intake to Reliable Delivery
A practical guide to orchestrating AI agents for software implementation with clear workflows, control points, and measurable outcomes.
Engineering5 min read
Operational Guardrails for AI Coding Agents: Best Practices, Caveats, and Failure Modes
How to keep agent-driven delivery safe and predictable with branching policy, review gates, and recovery patterns.
Enterprise AI5 min read
AI Dev Economics 2026: Pricing vs Value Across OpenAI, Anthropic, Google, GitHub Copilot, and xAI
A practical pricing and ROI lens for choosing model providers in agentic software delivery.
Engineering4 min read
The Operator's Dashboard: Building Ops Tools That Don't Need a Manual
The best ops dashboard is one your non-technical team can use without a Slack message to engineering. Here's what we built for ArchySocial's admin console redesign, and the principles that made the difference.
Engineering4 min read
Real-Time AI Video: Building Progress Tracking That Doesn't Gaslight Your Users
Real-time progress bars for AI video generation sound simple. Build one correctly \u2014 with job recovery, stale state handling, and honest progress reporting \u2014 and you'll appreciate why most apps don't bother.
Engineering4 min read
LinkedIn at Scale: Organisation Analytics, Follower Mentions, and Moderation Done Right
Organisation analytics, follower mentions, and comment moderation — the LinkedIn API features that matter for enterprise teams, and how ArchySocial surfaces them without burying users in raw API responses.
Engineering4 min read
One Post, Three Platforms: Expanding Social Publishing to Meta's Ecosystem
How we extended ArchySocial's publishing engine to support Facebook, Instagram, and Threads — and what the Meta API ecosystem taught us about building for multiple networks without multiplying your complexity.
Engineering5 min read
MCP OAuth 2.1: Why We Stopped Shipping API Keys and Let the Protocol Handle It
When VS Code Copilot asks for your MCP server connection, the user experience should be "click and go" — not "copy-paste this 48-character key, don't lose it, and definitely don't commit it to git." Here's how we implemented OAuth 2.1 + PKCE for ArchySocial's MCP server, and what the experience looks like from both sides of the handshake.