Engineering, research, and product writing on building production multi-agent systems with Tutti AI.https://tutti-ai.comen-GBhttps://tutti-ai.com/blog/what-an-agent-needs-after-it-workshttps://tutti-ai.com/blog/what-an-agent-needs-after-it-worksEvery framework can produce a working demo. The hard part is everything after — running on a schedule, getting better at the same task, and not billing you while idle. Three operational primitives that close the gap, and why the framework — not the user — should own them.Sun, 24 May 2026 00:00:00 GMTEngineeringdesignschedulingskillsserverlessChihabhttps://tutti-ai.com/blog/inbox-every-channel-one-agenthttps://tutti-ai.com/blog/inbox-every-channel-one-agentv0.25 lands inbound messaging on five platforms at once — Telegram, Slack, Discord, email, WhatsApp. The hard part wasn't writing five adapters. It was finding the one shape that made all five behave like the same thing.Sun, 17 May 2026 00:00:00 GMTEngineeringinboxdesignplatformsvoicesChihabhttps://tutti-ai.com/blog/smart-routing-the-design-storyhttps://tutti-ai.com/blog/smart-routing-the-design-storyMost agent turns don't need Opus. The problem is figuring out which ones do — at runtime, before you've spent the money. Here's how Tutti's SmartProvider does it, what we got wrong on the first try, and where it still falls short.Sun, 10 May 2026 00:00:00 GMTEngineeringsmart-routingcostrouterdesignChihabhttps://tutti-ai.com/blog/comparing-tutti-langgraph-crewai-autogen-mastrahttps://tutti-ai.com/blog/comparing-tutti-langgraph-crewai-autogen-mastraAn honest comparison. Each framework gets something right that Tutti doesn't, and vice versa. Here's where we differ, and where we'd recommend using something else.Sun, 03 May 2026 00:00:00 GMTProductcomparisonlanggraphcrewaiautogenmastraChihabhttps://tutti-ai.com/blog/opentelemetry-day-onehttps://tutti-ai.com/blog/opentelemetry-day-oneIf you can't explain what an agent did six hours after it did it, you can't operate it. Tutti emits OTEL spans for every run, LLM call, and tool invocation — and you wire it into whatever you already use.Sun, 26 Apr 2026 00:00:00 GMTEngineeringobservabilityOTELtelemetryChihabhttps://tutti-ai.com/blog/hitl-by-default-not-opt-inhttps://tutti-ai.com/blog/hitl-by-default-not-opt-inMost frameworks make human-in-the-loop something you wire in. By the time you remember to add it, the agent has already shipped a tweet. Tutti gates destructive operations by default — the wiring is the runtime.Sun, 19 Apr 2026 00:00:00 GMTEngineeringHITLdestructivesafetyChihabhttps://tutti-ai.com/blog/mcp-bridge-zero-gluehttps://tutti-ai.com/blog/mcp-bridge-zero-glueEvery framework eventually needs to talk to every other tool. Re-implementing every integration as a first-party voice is the wrong move. The MCP voice means any MCP server is one install away.Sun, 12 Apr 2026 00:00:00 GMTProductmcpvoicesextensibilityChihabhttps://tutti-ai.com/blog/prompt-injection-architecturehttps://tutti-ai.com/blog/prompt-injection-architectureIf you wait until the model has seen the malicious string, you've already lost. The defence has to be in the architecture: typed tool outputs, sanitised content, boundary markers, runtime isolation.Sun, 05 Apr 2026 00:00:00 GMTEngineeringsecurityprompt-injectionPromptGuardChihabhttps://tutti-ai.com/blog/permission-scopes-runtime-enforcedhttps://tutti-ai.com/blog/permission-scopes-runtime-enforcedMost frameworks document a permission model. Tutti enforces one. The difference is the runtime refuses to start when an agent has voices it didn't grant.Sun, 29 Mar 2026 00:00:00 GMTEngineeringsecuritypermissionsdesignChihabhttps://tutti-ai.com/blog/voices-the-plugin-modelhttps://tutti-ai.com/blog/voices-the-plugin-modelMost frameworks treat tools as a flat list. Every team rebuilds the same GitHub wrapper, the same Slack wrapper, the same Stripe wrapper. Voices are how Tutti stops that.Sun, 22 Mar 2026 00:00:00 GMTEngineeringvoicespluginsdesignChihabhttps://tutti-ai.com/blog/score-files-vs-graphshttps://tutti-ai.com/blog/score-files-vs-graphsAgent systems aren't general programs. They're a small set of repeating patterns. When you write them as code, the configuration disappears. When you write them as config, the system shape is the file.Sun, 15 Mar 2026 00:00:00 GMTEngineeringscoreconfigdesignChihabhttps://tutti-ai.com/blog/why-im-building-tuttihttps://tutti-ai.com/blog/why-im-building-tuttiI spent early 2026 trying to ship a real multi-agent workflow on top of every framework that almost worked. Each had a different deal-breaker. So I started writing the one I wished existed.Sun, 08 Mar 2026 00:00:00 GMTCompanymanifestooriginChihab