~/talk

Lessons From The Trenches: Building Rube

links:none on file
slides:none on file

Rube is a universal MCP server that connects LLMs to 500+ apps and manages huge context by writing bulky tool responses to a machine sandbox instead of embedding them in the prompt. It parallelizes independent calls with a multi-execute planner, stores remembered facts in a vector DB, and runs generated code in a remote workbench to handle complex cross-app flows. It handles OAuth onboarding, claims SOC2-level security, and demonstrates working end-to-end on real GitHub/Gmail workflows.

AI-written summary · about

  1. 0:00Introduction and audience setup
  2. 1:06Challenges with current MCP servers
  3. 3:57Wish list for an ideal MCP server
  4. 5:01Lessons learned from building features
  5. 6:06Solving large context window issues
  6. 7:30Parallel tool execution optimization
  7. 8:11Implementing memory for efficiency
  8. 9:16Handling complex user queries
  9. 10:20Demo of Rube in action
  10. 14:47What is Rube and why it exists
  11. 15:29Live demo with audience use case
  12. 23:13Comparing Rube with MCP gateway
  13. 23:54Showcasing the thank you slide
  14. 25:00Testing Rube with GitHub PR use case
  15. 30:13Summary and call to action
~/search
type to search 6,518 conferences, 143,918 speakers and 192,133 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
← →
on a talk: 5 seconds back / forward
0–9
on a talk: jump to 0%–90% of the recording
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.