vidtheque.

read-only demo

Videos -jY2T2PiJBE

MCP Apps: Extending the Frontier — Ido Salomon & Liad Yosef

index_state ready data_status ok

AI Engineer· published 2026-08-02· 0:18:38· en-US· indexed 2026-08-10 19:36

Open on YouTube

Scene timeline

  1. Shot 0, 0:00 to 0:03, 1 of 1 keyframes kept
  2. Shot 1, 0:03 to 0:05, 1 of 1 keyframes kept
  3. Shot 2, 0:05 to 0:12, 1 of 1 keyframes kept
  4. Shot 3, 0:12 to 0:16, 1 of 1 keyframes kept
  5. Shot 4, 0:16 to 0:17, 1 of 1 keyframes kept
  6. Shot 5, 0:17 to 0:44, 1 of 1 keyframes kept
  7. Shot 6, 0:44 to 0:51, 1 of 1 keyframes kept
  8. Shot 7, 0:51 to 0:52, 1 of 1 keyframes kept
  9. Shot 8, 0:52 to 0:56, 1 of 1 keyframes kept
  10. Shot 9, 0:56 to 0:59, 1 of 1 keyframes kept
  11. Shot 10, 0:59 to 1:14, 1 of 1 keyframes kept
  12. Shot 11, 1:14 to 1:28, 1 of 1 keyframes kept
  13. Shot 12, 1:28 to 1:41, 1 of 1 keyframes kept
  14. Shot 13, 1:41 to 1:51, 1 of 1 keyframes kept
  15. Shot 14, 1:51 to 2:21, 1 of 1 keyframes kept
  16. Shot 15, 2:21 to 2:28, 1 of 1 keyframes kept
  17. Shot 16, 2:28 to 2:31, 1 of 1 keyframes kept
  18. Shot 17, 2:31 to 2:46, 1 of 1 keyframes kept
  19. Shot 18, 2:46 to 2:59, 1 of 1 keyframes kept
  20. Shot 19, 2:59 to 3:12, 1 of 1 keyframes kept
  21. Shot 20, 3:12 to 3:43, 1 of 1 keyframes kept
  22. Shot 21, 3:43 to 3:45, 1 of 1 keyframes kept
  23. Shot 22, 3:45 to 4:06, 1 of 1 keyframes kept
  24. Shot 23, 4:06 to 4:11, 1 of 1 keyframes kept
  25. Shot 24, 4:11 to 4:14, 0 of 1 keyframes kept
  26. Shot 25, 4:14 to 4:17, 1 of 1 keyframes kept
  27. Shot 26, 4:17 to 4:21, 1 of 1 keyframes kept
  28. Shot 27, 4:21 to 4:22, 1 of 1 keyframes kept
  29. Shot 28, 4:22 to 4:25, 1 of 1 keyframes kept
  30. Shot 29, 4:25 to 4:27, 1 of 1 keyframes kept
  31. Shot 30, 4:27 to 4:37, 1 of 1 keyframes kept
  32. Shot 31, 4:37 to 4:56, 1 of 1 keyframes kept
  33. Shot 32, 4:56 to 5:03, 1 of 1 keyframes kept
  34. Shot 33, 5:03 to 5:33, 1 of 1 keyframes kept
  35. Shot 34, 5:33 to 6:02, 1 of 1 keyframes kept
  36. Shot 35, 6:02 to 6:32, 1 of 1 keyframes kept
  37. Shot 36, 6:32 to 6:41, 1 of 1 keyframes kept
  38. Shot 37, 6:41 to 7:21, 1 of 1 keyframes kept
  39. Shot 38, 7:21 to 7:23, 0 of 1 keyframes kept
  40. Shot 39, 7:23 to 7:30, 1 of 1 keyframes kept
  41. Shot 40, 7:30 to 7:55, 1 of 1 keyframes kept
  42. Shot 41, 7:55 to 7:59, 0 of 1 keyframes kept
  43. Shot 42, 7:59 to 8:32, 0 of 1 keyframes kept
  44. Shot 43, 8:32 to 8:35, 0 of 1 keyframes kept
  45. Shot 44, 8:35 to 8:50, 0 of 1 keyframes kept
  46. Shot 45, 8:50 to 8:54, 1 of 1 keyframes kept
  47. Shot 46, 8:54 to 9:26, 1 of 1 keyframes kept
  48. Shot 47, 9:26 to 9:59, 1 of 1 keyframes kept
  49. Shot 48, 9:59 to 10:32, 0 of 1 keyframes kept
  50. Shot 49, 10:32 to 10:51, 1 of 1 keyframes kept
  51. Shot 50, 10:51 to 10:52, 1 of 1 keyframes kept
  52. Shot 51, 10:52 to 10:56, 1 of 1 keyframes kept
  53. Shot 52, 10:56 to 10:57, 1 of 1 keyframes kept
  54. Shot 53, 10:57 to 11:07, 1 of 1 keyframes kept
  55. Shot 54, 11:07 to 11:25, 1 of 1 keyframes kept
  56. Shot 55, 11:25 to 12:03, 1 of 1 keyframes kept
  57. Shot 56, 12:03 to 12:25, 1 of 1 keyframes kept
  58. Shot 57, 12:25 to 12:28, 1 of 1 keyframes kept
  59. Shot 58, 12:28 to 12:36, 1 of 1 keyframes kept
  60. Shot 59, 12:36 to 12:46, 1 of 1 keyframes kept
  61. Shot 60, 12:46 to 13:03, 1 of 1 keyframes kept
  62. Shot 61, 13:03 to 13:12, 1 of 1 keyframes kept
  63. Shot 62, 13:12 to 13:22, 1 of 1 keyframes kept
  64. Shot 63, 13:22 to 13:39, 1 of 1 keyframes kept
  65. Shot 64, 13:39 to 13:49, 1 of 1 keyframes kept
  66. Shot 65, 13:49 to 14:09, 1 of 1 keyframes kept
  67. Shot 66, 14:09 to 14:14, 0 of 1 keyframes kept
  68. Shot 67, 14:14 to 14:21, 1 of 1 keyframes kept
  69. Shot 68, 14:21 to 14:37, 1 of 1 keyframes kept
  70. Shot 69, 14:37 to 14:52, 0 of 1 keyframes kept
  71. Shot 70, 14:52 to 14:59, 1 of 1 keyframes kept
  72. Shot 71, 14:59 to 15:33, 1 of 1 keyframes kept
  73. Shot 72, 15:33 to 15:36, 1 of 1 keyframes kept
  74. Shot 73, 15:36 to 15:57, 1 of 1 keyframes kept
  75. Shot 74, 15:57 to 16:02, 1 of 1 keyframes kept
  76. Shot 75, 16:02 to 16:06, 1 of 1 keyframes kept
  77. Shot 76, 16:06 to 16:17, 1 of 1 keyframes kept
  78. Shot 77, 16:17 to 16:25, 1 of 1 keyframes kept
  79. Shot 78, 16:25 to 16:42, 1 of 1 keyframes kept
  80. Shot 79, 16:42 to 16:58, 1 of 1 keyframes kept
  81. Shot 80, 16:58 to 17:08, 1 of 1 keyframes kept
  82. Shot 81, 17:08 to 17:40, 1 of 1 keyframes kept
  83. Shot 82, 17:40 to 17:48, 1 of 1 keyframes kept
  84. Shot 83, 17:48 to 17:50, 1 of 1 keyframes kept
  85. Shot 84, 17:50 to 17:52, 1 of 1 keyframes kept
  86. Shot 85, 17:52 to 17:56, 1 of 1 keyframes kept
  87. Shot 86, 17:56 to 17:57, 1 of 1 keyframes kept
  88. Shot 87, 17:57 to 17:58, 1 of 1 keyframes kept
  89. Shot 88, 17:58 to 18:01, 1 of 1 keyframes kept
  90. Shot 89, 18:01 to 18:06, 0 of 1 keyframes kept
  91. Shot 90, 18:06 to 18:11, 1 of 1 keyframes kept
  92. Shot 91, 18:11 to 18:16, 1 of 1 keyframes kept
  93. Shot 92, 18:16 to 18:19, 1 of 1 keyframes kept
  94. Shot 93, 18:19 to 18:21, 1 of 1 keyframes kept
  95. Shot 94, 18:21 to 18:37, 0 of 1 keyframes kept

95 shot(s).

keyframes kept every frame deduplicated

What was stored

cues
227
whisperx 227
chunks
33
from 227 cues
keyframes
84
kept of 95 captured
frames with text
84
1,876 lines read
chapters
11
from the source metadata
keyframe bytes
11.8 MB
word timings on 227 cues

Provenance

Each pipeline stage, its state and the model that produced it
stage state model started took
fetch done — 2026-08-09 21:22 0s
stt done — 2026-08-09 00:38 27s
chunk done — 2026-08-09 00:39 0s
text_embed done — 2026-08-10 19:36 1s
keyframe done — 2026-08-09 00:39 2m 54s
ocr done — 2026-08-09 00:42 33s
frame_embed done — 2026-08-10 19:36 14s

Frames, and what the machine read

  • 0:02 #0 done2 line(s)

    shot 0·sharpness 459.9

    1. AlEngineer0.96
    2. World's Fair1.00

  • 0:03 #1 done2 line(s)

    shot 1·sharpness 662.1

    1. AlEngineer0.95
    2. World's Fair0.99

  • 0:10 #2 done24 line(s)

    shot 2·sharpness 2744.8

    1. LAB & PLATINUM SPONSORS0.99
    2. Amazon AGI Lab0.98
    3. ANTHROP\C1.00
    4. Google DeepMind1.00
    5. MINIMAX0.94
    6. OpenAI0.92
    7. Akamai1.00
    8. arize1.00
    9. aws1.00
    10. Braintrust bright data0.98
    11. B1.00
    12. Browserbase1.00
    13. docker1.00
    14. :neo4j0.92
    15. ORACLE1.00
    16. PayPal1.00
    17. qodo1.00
    18. reducto1.00
    19. Sonar1.00
    20. Makers of1.00
    21. togetherai1.00
    22. Unblocked1.00
    23. WorkOS1.00
    24. SonarQube1.00

  • 0:15 #3 done2 line(s)

    shot 3·sharpness 170.2

    1. World's Fair1.00
    2. AlEngineer1.00

  • 0:17 #4 done6 line(s)

    shot 4·sharpness 1509.1

    1. AlEngineer0.98
    2. World's Fair0.92
    3. We built this talk yesterday1.00
    4. (so it might be outdated)1.00
    5. World's Fair0.96
    6. Engineering the future of Al0.99

  • 0:23 #5 done13 line(s)

    shot 5·sharpness 2499.7

    1. AlEngineer0.98
    2. World'sFair1.00
    3. PRESENTED BY1.00
    4. Microsoft1.00
    5. Ido Salomon0.99
    6. Creator @ MCP-UI1.00
    7. Co-Creator & Maintainer @ MCP Apps1.00
    8. Creator @ AgentCraft1.00
    9. x@idosall0.97
    10. idosal1.00
    11. in ido-salomon0.99
    12. World's Fair0.90
    13. Engineering the future of Al1.00

  • 0:45 #6 done12 line(s)

    shot 6·sharpness 2235.5

    1. AlEngineer0.97
    2. World'sFair0.97
    3. How can I help you0.95
    4. this morning?0.99
    5. PRESENTED BY1.00
    6. Microsoft1.00
    7. Our platform's activation rate has jumped1.00
    8. significantly over the past few days.0.99
    9. What's driving it?1.00
    10. 个0.61
    11. Worid's Fair0.93
    12. Engineering the future of Al0.99

  • 0:52 #7 done18 line(s)

    shot 7·sharpness 2766.0

    1. AlEngineer0.98
    2. World'sFair1.00
    3. Feb 281.00
    4. Mar 20.91
    5. Mar 40.94
    6. Mar61.00
    7. Mar 80.93
    8. Mar 100.95
    9. New onboarding cohort0.98
    10. Baseline1.00
    11. PRESENTED BY1.00
    12. Daily, Last 10 days | Amplitude 2.00.97
    13. Microsoft1.00
    14. Found it. A new onboarding flow was deployed on0.99
    15. Monday that has already improved activations by1.00
    16. 40%. Early results look strong!0.99
    17. TRACK 8· JULY 1, 20260.95
    18. Context Engineering1.00

  • 0:55 #8 done27 line(s)

    shot 8·sharpness 1694.0

    1. A0.99
    2. Amplitude1.00
    3. AlEngineer0.99
    4. World'sFair1.00
    5. Activation rate1.00
    6. < >0.79
    7. 40%1.00
    8. Mar 101.00
    9. 31.4%1.00
    10. PRESENTED BY1.00
    11. 30%1.00
    12. 19.2%1.00
    13. Microsoft1.00
    14. 20%1.00
    15. 10%1.00
    16. Feb 280.99
    17. Mar 20.96
    18. Mar 40.99
    19. Mar 60.94
    20. Mar 80.93
    21. Mar 100.96
    22. New onboarding cohort0.99
    23. Baseline1.00
    24. Daily, Last 10 days | Amplitude 2.00.98
    25. World's Fair0.99
    26. TRACK 8 • JULY 1, 20260.94
    27. Context Engineering1.00

  • 0:59 #9 done14 line(s)

    shot 9·sharpness 3033.5

    1. Daily, Last 10 days | Amplitude 2.00.97
    2. AlEngineer0.97
    3. World'sFair1.00
    4. Found it. A new onboarding flow was deployed on0.99
    5. Monday that has already improved activations by0.99
    6. 40%. Early results look strong!1.00
    7. PRESENTED BY1.00
    8. Microsoft1.00
    9. Makes sense! Can you show me where the1.00
    10. old flow was losing people?1.00
    11. ↑0.63
    12. World'sFair0.93
    13. TRACK 8· JULY 1, 20260.93
    14. Context Engineering0.99

  • 1:12 #10 done8 line(s)

    shot 10·sharpness 1148.1

    1. AlEngineer0.98
    2. We're used to text, but -0.96
    3. World'sFair1.00
    4. PRESENTED BY1.00
    5. Microsoft1.00
    6. World's Fair0.96
    7. TRACK 8· JULY 1, 20260.95
    8. Context Engineering1.00

  • 1:27 #11 done7 line(s)

    shot 11·sharpness 1654.4

    1. AlEngineer0.99
    2. We're used to text, but -0.99
    3. World'sFair1.00
    4. Text is the worst0.99
    5. World's Fair0.95
    6. TRACK 8· JULY 1, 20260.95
    7. Context Engineering1.00

  • 1:34 #12 done28 line(s)

    shot 12·sharpness 2339.7

    1. AlEngineer0.97
    2. The UX Gap0.96
    3. World'sFair1.00
    4. These are the full profiles for the three (3) members the new hire will work with. I am sharing0.99
    5. them all together so you can analyze faster. Each new member profile starts with Team0.99
    6. Member X, where X is the number (X = 1, 2, 3, etc.). This is also the last data set, so you can0.98
    7. analyze and give me the ideal candidate profile for our new hire.1.00
    8. Team Member 1 - Junior Java Developer0.98
    9. Role within the team: Responsible for developing, testing, and maintaining Java-based1.00
    10. applications.1.00
    11. Specific job duties and responsibilities: Write and test Java code, troubleshoot issues,1.00
    12. collaborate with other team members on projects, and participate in code reviews.0.99
    13. Skills and qualifications required for the job: Strong knowledge of Java, experience with0.99
    14. coding and debugging, familiarity with software development lifecycle (SDLC)0.99
    15. methodologies, knowledge of web technologies (HTML, CSS, JavaScript), and experience1.00
    16. with relational databases.1.00
    17. Experience level: 8 months with the company, fresh out of college.0.99
    18. Diversity & Demographic data: Male, 25 years old, Asian.1.00
    19. Training and development needs: Training on advanced Java concepts, software testing0.99
    20. methodologies, and project management.1.00
    21. Challenges with other members: May need guidance and support from senior team0.99
    22. members on more complex1.00
    23. < Share0.99
    24. Regenerate response1.00
    25. riers due to language and0.99
    26. World'sFair0.97
    27. TRACK 8· JULY 1, 20260.94
    28. Context Engineering0.99

  • 1:48 #13 done7 line(s)

    shot 13·sharpness 1366.1

    1. AlEngineer0.98
    2. World'sFair0.97
    3. What if every app in the world1.00
    4. could send its own Ul to the chat?1.00
    5. World's Fair0.95
    6. TRACK 8· JULY 1, 20260.95
    7. Context Engineering1.00

  • 1:54 #14 done10 line(s)

    shot 14·sharpness 1638.2

    1. Please show me my entries1.00
    2. AlEngineer0.99
    3. Of course! Here is the entry for today:1.00
    4. World'sFair1.00
    5. Reply to Claude...1.00
    6. @ Research0.91
    7. Sonnet 4.51.00
    8. World's Fair0.95
    9. TRACK 8 · JULY 1, 20260.93
    10. Context Engineering1.00

  • 2:24 #15 done6 line(s)

    shot 15·sharpness 831.1

    1. AlEngineer0.98
    2. World'sFair1.00
    3. Using MCP-UI0.97
    4. Worid's Fair0.94
    5. TRACK 8 · JULY 1, 20260.92
    6. Context Engineering1.00

  • 2:30 #16 done42 line(s)

    shot 16·sharpness 2777.5

    1. Ido Salomon1.00
    2. @idosal1 ·May 160.92
    3. Promote1.00
    4. Introducing MCP UI, the SDK that wires UI components directly into the0.99
    5. AlEngineer0.98
    6. MCP protocol1.00
    7. World's Fair0.99
    8. Watch the video to see the new phase in Al interaction!0.99
    9. Why browse when you can have universal apps that deliver the right Ul to0.99
    10. consume any data and perform any action?1.00
    11. Replace cumbersome walls of text with slick interactive components,0.99
    12. Wild Goose Case1.00
    13. bringing rich web experiences to MCP-enabled agents such as Claude and0.99
    14. THE FUTURE OF AGENTIC0.99
    15. Cursor.1.00
    16. INTERFACES1.00
    17. Let's build Jarvis so @kentcdodds doesn't have to!>>0.98
    18. AT E , 20250.55
    19. IMPM E //AMPT0.62
    20. linktr.ee/blockopensource0.85
    21. LOCATION0.99
    22. Bob1.00
    23. Performance Overview1.00
    24. Average 1250.99
    25. Reach 91%1.00
    26. LIAD YOSEF0.98
    27. 100 SALOMON0.96
    28. Feb0.99
    29. Mar1.00
    30. Apr0.99
    31. May1.00
    32. Jun1.00
    33. 12431.00
    34. 2891.00
    35. 671.00
    36. Pending1.00
    37. Blocked1.00
    38. 0:041.00
    39. end a message0.99
    40. World's Fair0.98
    41. TRACK 8· JULY 1,20260.96
    42. Context Engineering0.98

  • 2:41 #17 done11 line(s)

    shot 17·sharpness 1335.2

    1. AlEngineer0.98
    2. World'sFair1.00
    3. An open protocol for1.00
    4. interactive apps over0.99
    5. MCP standardizing UI1.00
    6. transmission and1.00
    7. Ul<>Host0.95
    8. communication1.00
    9. World's Fair0.96
    10. TRACK 8· JULY 1, 20260.95
    11. Context Engineering1.00

  • 2:56 #18 done6 line(s)

    shot 18·sharpness 1188.4

    1. AlEngineer0.99
    2. World'sFair1.00
    3. AI0.99
    4. World's Fair0.99
    5. TRACK 8• JULY 1, 20260.96
    6. Context Engineering0.98

  • 3:09 #19 done19 line(s)

    shot 19·sharpness 1974.3

    1. Home » Posts0.95
    2. MCP Apps - Bringing UI Capabilities To0.99
    3. AlEngineer0.92
    4. World'sFair1.00
    5. MCP Clients0.95
    6. MCP Apps is now live as the first official MCP extension — tools can return interactive UI0.99
    7. components that render directly in the conversation.0.97
    8. January 26, 2026 - 9 min - MCP Core Maintainers0.98
    9. Table of Contents0.97
    10. Today, we're announcing that MCP Apps are now live as an official MCP extension.0.99
    11. Tools can now return interactive Ul components that render directly in the conversation:0.99
    12. dashboards, forms, visualizations, multi-step workflows, and more. This is the first official1.00
    13. MCP extension, and it's ready for production.1.00
    14. Watch video on YouTube1.00
    15. Video player configuration error0.97
    16. Error1531.00
    17. World'sFair0.95
    18. TRACK 8·JULY 1,20260.97
    19. Context Engineering0.98

  • 3:33 #20 done11 line(s)

    shot 20·sharpness 2460.7

    1. World'sFair1.00
    2. AlEngineer0.98
    3. TheFork1.00
    4. .monday.com0.99
    5. IIElevenLabs0.97
    6. POSTMAN0.99
    7. shopify1.00
    8. obot1.00
    9. Worid's Fair0.95
    10. TRACK 8· JULY 1, 20260.96
    11. Context Engineering1.00

  • 3:45 #21 done5 line(s)

    shot 21·sharpness 657.0

    1. AlEngineer0.99
    2. World's Fair0.97
    3. World's Fair0.95
    4. TRACK 8 · JULY 1, 20260.93
    5. Context Engineering1.00

  • 4:02 #22 done30 line(s)

    shot 22·sharpness 3554.7

    1. + Connect0.92
    2. AlEngineer0.99
    3. Sam Morrow· 2nd0.98
    4. World'sFair1.00
    5. Two ne1.00
    6. MCP Apps compatibility in ChatGPT0.99
    7. - we sh0.93
    8. Build portable MCP Apps UIs that run in ChatGPT0.98
    9. - we sh0.99
    10. Overview1.00
    11. Try it 00.95
    12. ChatGPT supports the MCP Apps open standard for embedded app UIs.0.99
    13. Should1.00
    14. MCP Apps Uls run inside an iframe and communicate with the host over a standard bridge (ui/* JSON-RPC over0.98
    15. Ran Get my0.97
    16. postMessage). ChatGPT implements this same iframe-and-bridge model, so you can build your Ul once and run it in0.98
    17. ChatGPT and other MCP Apps-compatible hosts.0.99
    18. Existing Apps SDK APls remain supported, and new, experimental capabilities ship first in the Apps SDK. OpenAl helped0.99
    19. shape the MCP Apps standard from ChatGPT Apps, and new capabilities move into the MCP spec after shape and0.99
    20. e0.68
    21. functionality validation.0.98
    22. Build with the MCP Apps standard keys and bridge by default. Use window. openai when you need ChatGPT-specific0.99
    23. capabilities.1.00
    24. CAaron Koivunen and 59 others0.99
    25. 13 comments - 6 reposts0.96
    26. 451.00
    27. h 6.3K0.81
    28. World's Fair0.99
    29. TRACK 8·JULY 1,20260.97
    30. ContextEngineering1.00

  • 4:07 #23 done40 line(s)

    shot 23·sharpness 3449.5

    1. Ruben Casas1.00
    2. AlEngineer0.98
    3. Sam Morrow · 2nd0.97
    4. @Infoxicador1.00
    5. World's Fair0.98
    6. + CitLlub0.78
    7. Announcing Official MCP Apps Support in Postman!0.98
    8. Two ne1.00
    9. MCP Apps compatibil1.00
    10. You can now debug and test MCP servers that implement the new0.99
    11. specification, inspect requests, responses, and events, and validate0.99
    12. - we sh0.93
    13. Build portable MCP Apps UIls that ru0.95
    14. interactive behaviour defined by MCP Apps.0.98
    15. - we sh0.94
    16. Overview1.00
    17. Try it o0.92
    18. ChatGPT supports the MCP Apps oper0.98
    19. Should1.00
    20. MCP Apps Uls run inside an iframe anc0.99
    21. in1.00
    22. Ran Get my0.98
    23. postMessage). ChatGPT implements0.98
    24. ChatGPT and other MCP Apps-comp1.00
    25. helped0.96
    26. Existing Apps SDK APIs remain supp0.99
    27. shape the MCP Apps standard from0.99
    28. e0.93
    29. functionality validation.0.99
    30. ecific1.00
    31. Build with the MCP Apps standard0.99
    32. capabilities.0.98
    33. CAaron Koivunen and 59 others0.98
    34. 10:00 AM - Jan 9, 2026-20.8KViews0.96
    35. 0:10/0:410.95
    36. à0.64
    37. ↑0.63
    38. World's Fair0.98
    39. TRACK 8· JULY 1, 20260.93
    40. Context Engineering1.00

Transcript

227 cues· 3,104 words· 16,094 chars

  1. 0:12 Hi.
  2. 0:13 So hi, everyone.
  3. 0:15 We did this talk yesterday, so it might be out of date.
  4. 0:18 I'm Ido Salomon.
  5. 0:20 I am the creator of MCPUI and co-creator and maintainer of MCP apps in the MCP steering committee.
  6. 0:25 I also created AgentCraft, if you were in the talk yesterday.
  7. 0:30 I'm Liad, I work with the DAW and MCPUI.
  8. 0:32 I'm also the co-creator and maintainer of the MCP apps spec and recently co-founded Aura, which is a research lab for the agentic web.
  9. 0:40 And we're going to talk a little bit more about it later.
  10. 0:45 So MCP apps are all around us.
  11. 0:47 You might not even realize it, but all the fancy apps you have today in ChatGPT and VS Code and Slack are actually all based on MCP and the MCP app spec.
  12. 1:01 and if we take a step back and we ask why do we need mcp apps what's the idea behind mcp or mcp apps so when we work with chats with chat clients we used to text because that's the natural interface but text is really the worst way to convey a lot of information right because we don't want walls of text and actually this is the main blocker from companies to build an mcp server
  13. 1:25 they don't want to be reduced to a textual database.
  14. 1:27 They don't want to lose their brand identity in the process.
  15. 1:30 They don't want their data that they worked so hard on building the UX for to look something like this.
  16. 1:37 So instead of this, what if the apps could just send their UI to the chat, right?
  17. 1:46 What if every service and every brand could just send their user interface to the chat?
  18. 1:51 So instead of us looking at something like this,
  19. 1:54 we could just have the apps send their own identity, their own UI chunks into the chat.
  20. 2:00 And then we take a look and we say, OK, yeah, I know this is Shopify in the middle.
  21. 2:03 I know this is Hugging Face.
  22. 2:05 I know this is Monday.
  23. 2:06 And what if we don't want to do it only as a visualization?
  24. 2:10 We also want to do it interactive.
  25. 2:12 So we want the users to be able to actually interact with Hugging Face, for example, and for Hugging Face to actually do something with it.
  26. 2:22 So we don't have to imagine the future, as we said, with MCP UI, which I created in May last year, and took that, which is essentially like an open protocol for interactive applications over MCP.
  27. 2:38 So it's not only how you transmit UI, but also how that UI, that application, communicates with the host.
  28. 2:47 And just a few months ago, we partnered with Anthropic and OpenAI to create the official extension to MCP, which we call MCP Apps, based on MCP UI, MCP SDK, and other solutions in the field.
  29. 3:01 The launch was pretty cool, with Cloud and VS Code supporting it to begin with.
  30. 3:06 But now, obviously, also OpenAI and others have adopted it.
  31. 3:13 Yeah, and there are a lot of early adopters to MCP UI.
  32. 3:17 Eleven Labs, Shopify, Postman, those are one of the first companies to support it.
  33. 3:22 Back like a year ago, they were the one believing in this spec, in this vision.
  34. 3:27 And Goose also supported it.
  35. 3:30 And it's a funny anecdote, because today Block released their agent e-commerce solution that is based on MCP apps.
  36. 3:36 So a year ago, Goose was the first client to support MCP UI, and now it is part of Block's
  37. 3:45 And today we have a lot more clients supporting MCP UI.
  38. 3:49 We have Cursor and we have Copilot and GitHub.
  39. 3:55 ChatGPT support MCP apps.
  40. 3:57 ChatGPT apps that you know are actually based on MCP apps.
  41. 4:01 And OpenAI actually recommend using
  42. 4:04 MCP apps as a protocol to build Church GPT apps, Postman, and a lot more.
  43. 4:10 And obviously, Cloud supports MCP apps.
  44. 4:12 But you also have a large community around it.
  45. 4:15 So people start to build plug-ins to MCP apps and integrations to different agents and also courses on how to build MCP apps.
  46. 4:24 This is Pi integration for MCP apps.
  47. 4:26 So we have a large community around it.
  48. 4:29 There's a repo, xApp, which is the repo for MCP apps, where everyone can just come and propose PRs and ideas of how to extend this spec.
  49. 4:39 And we have a work group in the MCP committee, and we're convening every three weeks.
  50. 4:43 We have a tri-weekly meeting on the future of the protocol and how to make the spec not just serve the big labs, but also the community.

Chapters

  1. 0:00 Why we need MCP Apps
  2. 1:52 From walls of text to interactive views
  3. 2:29 MCP UI, created and adopted
  4. 4:26 An open working group in the MCP committee
  5. 5:04 How a tool call becomes an interface
  6. 6:31 Standardizing the flow
  7. 8:52 The architecture: resources and web components
  8. 10:10 Consuming apps through the browser
  9. 14:29 What's still evolving in the spec
  10. 16:07 Interoperability across hosts
  11. 17:14 Write once, reach hundreds of millions

Open at this second