Give your AI the Power to Present.

MyLens is a presentation layer for all your AI tools. Instead of walls of text, your agents can now explain complex outputs visually, with narration and personalization.

Your AI

Assistant

Your AI + MyLens

Assistant

0:00

How to connect your AI to MyLens with MCP

Ask your AI to connect once. After that, it can use MyLens whenever a visual presentation is better than a text reply.

Copy and send this to your agent

Connect MyLens MCP at https://mcp.mylens.ai

If your AI cannot connect MCP itself, with our MCP URL.

Setup guides by platform

How it works

MyLens helps your AI communicate the way people understand best—narrated visuals personalized for each person, easier to grasp and share.

Effective AI communication

Delivered is not the same as understood.

Information creates value only when people can receive it, understand it, and act on it. MyLens makes that possible.

Without MyLens

Complex AI output

YouLong text. Same.

TeamLong text. Same.

StakeholdersLong text. Same.

With MyLens

Complex AI output

MyLens

YouPersonalized. Narrated. Visual

TeamPersonalized. Narrated. Visual

StakeholdersPersonalized. Narrated. Visual

It's not about what sounds good to the sender. It's about what the audience can understand.

If your AI has something complex to say, let it visually present to you.

Agent chat
Assistant
Agent chat
Assistant
Here is a full technical assessment of Northline’s current product stack and how presentation should fit into the agent workflow.
Architecture overview Northline runs a multi-tenant SaaS control plane with a TypeScript web client and Go services for long-running generation. Core API traffic uses authenticated REST plus SSE job streams. Persistent records live in Postgres; job state and rate limits sit in Redis; source files and pinned designs land in object storage.
Key surfaces today:
Workspace and storyboard models
Source library with provenance names
Theme / agent / visual skills
Share links returned after generation
Data plane rules Substantive research must be persisted as a source before visuals are planned. That keeps citations, narration, explain, and later edits aligned to the same material. Short creative prompts can skip sources, but memo-length analysis cannot.
What should always become a source:
Agent-authored research memos
Customer transcripts and call notes
Metric exports and anomaly tables
Competitive landscape drafts
MCP integration path MyLens MCP is a stateless gateway in front of Core. ChatGPT, Claude, and Cursor connect over streamable HTTP, authenticate with Auth0, and exchange tokens via On-Behalf-Of. Authorization stays in Core so plan limits cannot drift between the product UI and agent connectors.
Typical source-first flow:
`list_workspaces` to resolve the workspace slug
`add_source` with the finished analysis
`create_storyboard` for the presentation board
`add_sources_to_storyboard` to attach the source
`create_stories` with audience, goal, theme, and narration
Return the MyLens URL to the user
Auth and privacy defaults Unauthenticated MCP calls return 401 with a discovery challenge. Storyboards created by agents start private. Workspace sharing for members/guests is enabled after successful generation by default. Public `anyone` links require an explicit audience change.
Generation model `create_stories` starts a mastermind plan that fans out into child visual jobs. Tools wait on SSE progress rather than exposing raw polling. When requested, one run can produce:
Core visual structure
Pinned design treatment
Synchronized AI narration
Rendering and personalization Presentations are ordered storyboards, not chat dumps. Themes and agent skills steer color, typography, tone, and defaults from an unlimited theme gallery plus workspace-owned styles. The same underlying analysis can be reshaped for executives, technical teams, or customers by changing audience context and narrative depth.
Operational risks to watch
Plan-limit exhaustion during bursty agent traffic
Missing workspace membership after OAuth
Idle timeouts on long generation without progress heartbeats
Prompt-only facts that never entered the source library
Recommended next moves
Standardize a source-first MCP playbook for every research agent
Default vague presentation asks to canvas + pinned design + narration
Keep customer-facing boards on workspace sharing until explicit public publish
Treat MyLens as the presentation layer, not another markdown formatter
Bottom line: Northline’s agents already produce high-value analysis. The bottleneck is human understanding. A presentation layer through MCP turns that wall of text into a shareable, narrated visual brief leadership can act on.
Calling MyLensPresent for leadership · narrated
Agent chat
Assistant
Here is a full technical assessment of Northline’s current product stack and how presentation should fit into the agent workflow.
Architecture overview Northline runs a multi-tenant SaaS control plane with a TypeScript web client and Go services for long-running generation. Core API traffic uses authenticated REST plus SSE job streams. Persistent records live in Postgres; job state and rate limits sit in Redis; source files and pinned designs land in object storage.
Key surfaces today:
Workspace and storyboard models
Source library with provenance names
Theme / agent / visual skills
Share links returned after generation
Data plane rules Substantive research must be persisted as a source before visuals are planned. That keeps citations, narration, explain, and later edits aligned to the same material. Short creative prompts can skip sources, but memo-length analysis cannot.
What should always become a source:
Agent-authored research memos
Customer transcripts and call notes
Metric exports and anomaly tables
Competitive landscape drafts
MCP integration path MyLens MCP is a stateless gateway in front of Core. ChatGPT, Claude, and Cursor connect over streamable HTTP, authenticate with Auth0, and exchange tokens via On-Behalf-Of. Authorization stays in Core so plan limits cannot drift between the product UI and agent connectors.
Typical source-first flow:
`list_workspaces` to resolve the workspace slug
`add_source` with the finished analysis
`create_storyboard` for the presentation board
`add_sources_to_storyboard` to attach the source
`create_stories` with audience, goal, theme, and narration
Return the MyLens URL to the user
Auth and privacy defaults Unauthenticated MCP calls return 401 with a discovery challenge. Storyboards created by agents start private. Workspace sharing for members/guests is enabled after successful generation by default. Public `anyone` links require an explicit audience change.
Generation model `create_stories` starts a mastermind plan that fans out into child visual jobs. Tools wait on SSE progress rather than exposing raw polling. When requested, one run can produce:
Core visual structure
Pinned design treatment
Synchronized AI narration
Rendering and personalization Presentations are ordered storyboards, not chat dumps. Themes and agent skills steer color, typography, tone, and defaults from an unlimited theme gallery plus workspace-owned styles. The same underlying analysis can be reshaped for executives, technical teams, or customers by changing audience context and narrative depth.
Operational risks to watch
Plan-limit exhaustion during bursty agent traffic
Missing workspace membership after OAuth
Idle timeouts on long generation without progress heartbeats
Prompt-only facts that never entered the source library
Recommended next moves
Standardize a source-first MCP playbook for every research agent
Default vague presentation asks to canvas + pinned design + narration
Keep customer-facing boards on workspace sharing until explicit public publish
Treat MyLens as the presentation layer, not another markdown formatter
Bottom line: Northline’s agents already produce high-value analysis. The bottleneck is human understanding. A presentation layer through MCP turns that wall of text into a shareable, narrated visual brief leadership can act on.
MyLens completed
0:00

Agent chat

Assistant

MyLens gives your AI the ability to show, not just tell.

Ask ChatGPT, Claude, Cursor, Notion, and the rest to present the work visually to you. Complex chat is hard to understand. A visual walkthrough is much faster to digest, share, and stay engaging.

ChatGPT

You have a data analysis, finance comparison, or research thread in ChatGPT—use MyLens when the findings are too long to scan.

Claude

You have a long document, contract, or research synthesis in Claude—use MyLens when the argument is too dense to hold in your head.

Cursor

You have an architecture review, migration plan, or PR diagnosis in Cursor—use MyLens when the team needs the system, not the comment dump.

Claude Code

You have a codebase investigation, root cause, or refactor plan in Claude Code—use MyLens when the path and risks are hard to follow.

VS Code

You have debugging notes or an implementation plan in VS Code—use MyLens when you need to walk someone through the work.

Slack

You have a long Slack thread, incident channel, or standup recap—use MyLens when decisions, owners, and blockers are buried in messages.

Pi

You have a planning chat, decision comparison, or briefing in Pi—use MyLens when you want to send it to someone who was not in the conversation.

Gemini

You have a product comparison, plan, or research thread in Gemini—use MyLens when the findings span too many documents to reread.

Copilot

You have a meeting recap, business review, or action plan in Copilot—use MyLens when the team needs the outcome, not another memo.

Perplexity

You have sourced market research or a competitive landscape in Perplexity—use MyLens when the evidence is too long to reread.

Windsurf

You have a repo investigation or change-impact writeup in Windsurf—use MyLens when the flow and next steps are hard to follow.

Custom MCP agent

You have internal analytics, an ops report, or customer intelligence from a custom agent—use MyLens when that output is too dense as text.

Why visual

People communicate in four ways. Only one scales complexity.

When AI output gets dense and complex, a narrated personalized visual communication is the most effective channel people easily digest, remember, and share.

Four ways people communicate

Verbal

People talk.

Non-verbal

Body language and gesture.

Written

Documents, chat, and memos.

This is MyLens

Visual

Imagery, charts, symbols, and design.

Why visual wins

Remembered longer

People retain far more from visuals than from text or speech.

Universally understood

A clear visual crosses language, age, and background.

Simplifies complexity

Dense analysis becomes one graspable picture anyone can follow.

Instantly recognizable

Icons, shapes, and structure carry meaning at a glance.

More impactful

Visuals are engaging, eye-catching, and hold attention.

Straight to the point

The takeaway lands without wading through paragraphs.

Your AI already writes well. MyLens lets it show—in any theme, ready to present.

Frequently asked questions

MyLens gives any MCP-connected agent a way to answer with a narrated visual presentation instead of another wall of text. The result comes back as a shareable link in the same chat, so people can understand complex agent work in depth and personalized.

Any tool or custom agent that supports MCP, including ChatGPT, Claude, Claude Code, Cursor, VS Code, Notion, and internal or custom agents.

Ask your agent to connect to https://mcp.mylens.ai, or add that URL under Custom MCP / Custom Plugins. If the platform asks for a client ID, paste the one on this page—there is no client secret. Connect once and MyLens stays available whenever a visual presentation is better than a text reply. If you have a problem connecting, contact us and we will help soon.

Your AI agent can use MyLens whenever it decides an answer is complex and would be more effective to show visually. You can also ask it directly to use MyLens whenever you want something presented to you instead of explained in a long text response.

MyLens uses generative AI to create any visual, diagram, or chart you can imagine—not just a fixed set of templates. You can pair every visual with unlimited customized themes and styles, so the result can match your brand, audience, or personal taste and arrive presentation-ready.

Yes. Themes are not limited to a fixed gallery or simple color changes. MyLens can create a theme from your brand or generate an entirely original style as part of the AI visual itself, so the result can follow any aesthetic you imagine.

MyLens adapts the audience, theme, visual structure, level of detail, and information density to your goals, context, and existing knowledge. The result feels relevant and familiar to you, making complex ideas easier to digest and helping you make informed decisions.

Yes. Agents can request synchronized AI narration alongside the visual story, so the presentation explains itself to someone who was not in the original chat.

Yes. Every presentation is easy to share with teammates, clients, or stakeholders through a link.

Yes. You can connect your AI agent to MyLens via MCP and start creating narrated visual presentations for free. Upgrade when you need higher limits or advanced plan features.

Storyboards agents create start private to your workspace. Wider sharing with members or guests, or a public link, is an explicit choice—not the default.