Curious what others are building. Are you running multiple AI agents? What architectures work? What fails spectacularly?
Frameworks handle individual agent capabilities well. What they don't handle: preventing two agents from silently overwriting each other's work on shared state. It's a classic race condition but in AI systems the output looks reasonable, so you don't notice it until production.
We open-sourced a coordination layer that adds atomic state management to any framework (LangChain, AutoGen, CrewAI, MCP, etc.): https://github.com/Jovancoding/Network-AI
What works: role clarity + veto rights. One agent can only block, never propose. One agent makes calls, others can raise flags. You stop the chatbot parliament problem and actually get decisions.
The other pattern worth stealing from production systems: treat inbound events (emails, webhooks, form submissions) as the task boundary, not the conversation turn. An agent that owns a mailbox and processes messages one at a time is dramatically more auditable than one that's always-on and decides what to react to. You can replay it, diff its outputs, and understand why it did what it did.
It takes a plan, breaks it into dependent tasks, has human-in-the-loop for approval, and then is fire-and-forget after the plan is started with parallel agent workers. Has complete code review loops and testing loops for accuracy and quality. Idempotent retries and restarts...Completely frontend-driven so I don't have to deal with dumb terminals like claude code...
It's a blind fire n forget go worker danse.
wich can be hold as monitoreed or scale as multiple instances if needed by simple parameters.
Basicaly, It's a job as librairy patern.
If you dont need real time, its bulletproof and very llm friendly.
and a good token saver by the batching abilities.
Your queue is a struct with New(db) — it knows submit, poll, complete, fail, nothing else.
Your worker is another struct that loops on the queue and dispatches to handlers registered via RegisterHandler("type", fn). Your handlers are pure functions (ctx,payload) → (result, error) carried by a dependency struct.
Main just assembles: open DB, create queue, create worker, register handlers, call worker.Start(ctx). Result: each handler is unit-testable without the worker or network, the worker is reusable across any pipeline, and lifecycle is controlled by a simple context.Cancel().
Bonus: here the queue is a SQLite table with atomic poll (BEGIN IMMEDIATE), zero external infra.
The whole "framework" is 500 lines of readable Go, not an opaque DSL. TL;DR: every service is a library with New() + Start(ctx), the binary is just an assembler.
The "all in connectivity" pattern means every capability in your system — embeddings, document extraction, replication, MCP tools — is called through one interface: router.Call(ctx,"service", payload).
The router looks up a SQLite routes table to decide how to fulfill that call: in-memory function (local), HTTP POST (http), QUIC stream (quic), MCP tool (mcp), vector embedding (embed), DB replication (dbsync), or silent no-op (noop).
You code everything as local function calls — monolith. When you need to split a service out, you UPDATE one row in the routes table, the watcher picks it up via PRAGMA data_version, and the next call goes remote.
Zero code change, zero restart. Built-in circuit breaker, retry with backoff, fallback-to-local on remote failure, SSRF guard.
The caller never knows where the work happens.
That's the "job as library" pattern: the boundary between monolith and microservices is a config row, not an architecture decision.
I'm using Openclaw + Opus. Several subagents.
However, performance is degraded when using subagents - scraping is less smart, content is worse written, etc.
I'm curious about using different instances instead, but not sure how to use a shared memory foundation effectively.
Check out htpps://agentbus.org