Skip to main content

Dynamic Bundles: Hub-Style Power Inside Any Bundle

· 3 min read
Tony Lewis
MCPBundles
Ask AI:
PerplexityClaudeChatGPTGoogleGrok

Tool overload is real.

It shows up as lag. Wrong tool picks. Weird, half-finished workflows. Or the model just dumps a wall of raw data at you and calls it a day.

We’ve always had a simple answer: keep bundles focused. 5–15 tools for one job.

That still works great.

But sometimes you do want a big bundle. A real “everything I use for this role” bundle.

Now you can do that without turning your AI into a confused mess.

Every bundle can run in Dynamic.

Direct vs Dynamic (two modes, one bundle URL)​

Bundles now have an Execution Mode. Pick what fits.

  • Direct: classic. Your AI sees the full tool list and calls tools directly.
  • Dynamic: starts small, discovers tools only when needed, and runs bigger workflows without loading a giant tool list up front.

Same bundle URL. Same slug. Same credential bindings.

Different behavior.

Why Dynamic exists​

The two biggest scaling problems are:

  1. Tool definition bloat If a bundle contains 100+ tools, forcing an AI client to load every tool definition up front gets slow and fragile.

  2. Intermediate-result bloat Even if the model picks the right tool, big results can swamp the conversation and throw off everything that comes next.

Dynamic mode keeps both under control by:

  • discovering tools on-demand
  • executing workflows in a code execution environment
  • returning only the final output to the conversation

This is built on the same ideas Anthropic has been publishing lately:

The rule: Direct is limited to 100 tools​

Direct is still the best choice for small bundles.

But once a bundle goes past 100 tools, Direct becomes the wrong shape. Too much to load. Too much to reason over.

So we enforce a simple product constraint:

  • If a bundle has more than 100 tools, it must use Dynamic mode.

The bundle builder makes this obvious while you’re building, and the backend enforces it so you can’t end up in a weird “works for some clients, breaks for others” state.

Dynamic mode is “Hub-style,” but bundle-scoped​

If you’ve used the Hub, Dynamic will feel familiar—just scoped to one bundle.

Dynamic bundles start with a small surface:

  • Search tools in this bundle
  • Load tool info only when needed
  • Run dynamically (execute workflows, process results, return the answer)

That means:

  • You can build a large bundle and still keep it fast.
  • Credentials stay bundle-scoped (different bundles can use different connections).
  • Your AI client doesn’t get overwhelmed by thousands of tool definitions.

When to use what​

  • Direct: everyday work in one domain with a focused bundle.
  • Dynamic: larger bundles, multi-step workflows, or anytime you care about speed and reliability when the tool library is big.

Getting started​

  • Open any bundle → Settings → Execution Mode
  • Choose Dynamic (works at any bundle size)
  • If your bundle grows past 100 tools, Direct is no longer available for it

If you want background on how the Hub avoids overload and why search + code execution works, read:

Questions, edge cases, or “this is the annoying workflow I wish worked”… post it on Reddit.