Conversation Protocol for the AI Era [1]

[1] Shrinked standardizes conversational intelligence across all your sources. Bring your meetings, calls, interviews, and media to our context-aware storage with precise attribution — powering AI workflows and API processing.

UPLOAD
MP3
1M 8S

Upload hundreds of hours of audio and video content with a single click. Our protocol handles everything from corporate meetings to social media content, automatically categorizing and preparing all your conversational data for AI processing.

legal depositionsmeetingspodcastspolitical debatespublic hearingsyoutube
AUTOMATION
EVERY FRIDAY

Set up intelligent workflows that process your conversation library automatically. Our protocol converts all content into standardized JSON structures in docStore, enabling scheduled analysis and report generation without manual intervention.

quarterly reportsweeklysummariescontinuous contexttrend detectionscheduled
PROCESSING
CLAUDE-3-7-SONNET

Transform raw conversations into structured knowledge network with perfect citation and relationship mapping. Unlike simple summarization tools, we deliver comprehensive, interconnected insights spanning hundreds of pages, maintaining every contextual relationship.

CLAUDE 3.7 SONNETchapter extractionresearchsemantic linkingGemini 2.0
EXPORT TO EMAIL
PDF

Export intelegent reports with comprehensive attribution at any scale. From executive summaries to hundred-page research papers, all exports maintain perfect citations and can be accessed via email or our developer API.

jsonemail deliverypdf reportingmarkdown exportAPIteam distribution

The AI Contextualization Reimagined [2]

[2] Unlike traditional LLMs that obscure sources, our protocol maintains perfect attribution throughout. We convert conversations into structured JSON, connect related discussions across sources, and rebuild them into coherent knowledge where every insight traces back to its exact origin—revolutionizing conversational intelligence by prioritizing verifiability over black-box summaries.

[Conversation a]

/ real-time capture

Hey everyone, how's it going?
Not great actually.
Alex just pinged me about those dashboard performance issues we've been seeing.
It's loading at 4.7 seconds now!
That's way too slow.
Yikes, that's pretty bad.
Yeah, and we need this fixed before the client demo on Friday.
Have we figured out what's causing it?
The Product Development team thinks it's the database queries.
10 
Should we loop in the backend team?
11 
Let's wait and see what Product Dev comes back with first.

[Conversation b]

/ Team Follow-up Meeting

50 
Alright, I've got an update on the dashboard performance issue.
51 
Oh thank goodness. What did you find?
52 
So it turns out it wasn't the database after all.
53 
Lisa dug into it and found it's the image compression algorithm we implemented last month.
54 
Oh interesting! I wouldn't have guessed that.
55 
We've managed to reduce the load time to 2.3 seconds in our test environment.
56 
Just in time for the client demo tomorrow.
57 
That's a huge improvement! Are we documenting this somewhere?
58 
Definitely. Lisa's adding it to the technical docs now.
PERFORMANCE REPORT (289 P.)
The team identified critical dashboard performance issues with loading times of 4.7 seconds [4], initially attributed to database problems by the Product Development team [9]. This issue required urgent resolution before an upcoming client demo [7]. Later investigation by Lisa revealed the actual cause was the image compression algorithm [15]
Read the full report

RESEARCH

The API output includes all connections related to the image compression algorithm [15], linking conversations, server logs, and issue tracker entries for a comprehensive overview.

{"reference": "image compression algorithm",
"source_id": "conversation-b-feb-10","line": 53,"speaker": "Lisa Chen",
"connections": [
{"doc_id": "eng-weekly-jan-12",
"context": "Implementation planning for dashboard optimization",
"relevance_score": 0.87},
{"doc_id": "server-logs-feb-02",
"context": "Performance degradation analysis","relevance_score": 0.92},
{"doc_id": "issue-tracker-feb-05",
"context": "Bug report #4392: Dashboard loading time regression",
"relevance_score": 0.95}
]}

Source Analysis

[+]

Key Stakeholders

[+]

Related Conversations

[+]

Performance Metrics

[+]

Documentation

[+]

Control your source data[3]

[3] DocStore is the industry's first standardized conversational repository powered by our protocol. Create both public knowledge hubs and private repositories — with every insight maintaining perfect attribution to its source.

Shrinked AI Dashboard

Document Processing

Import any document format and automatically extract structured data with our advanced processing engine

Structured Storage

All data is organized into queryable fields, making information retrieval fast and precise

Access Controls

Set visibility permissions for each document to control what gets shared publicly or kept private

VisibilityCategoryTitleCreatedLogicTagsActions
Public
Social MediaElon Musk's Twitter Misinformation: A Case Study in Fact-CheckingJan 04Topic Analysis
MisinformationTwitterElon Musk
View
Private
Real EstateHome Inspection Reveals Major Issues with Deck and HVACOct 31Structured Extract
Home InspectionStructural IssuesHVAC
View
Public
PoliticsBernie Sanders on Politics, Healthcare, and Social ChangeOct 24Conversation Parse
Bernie SandersHealthcarePolitics
View
Private
BusinessManaging Partner's Document Workflow and PandaDoc ExperienceJan 17Dialog Extract
PandaDocDocument ManagementWorkflow
View
Public
FinanceNavigating the Venture Capital Landscape: Insights for FoundersOct 01Multi-Source Merge
Venture CapitalFundraisingStartup
View

Swipe horizontally to view the complete document table.