Markdown, published beautifully.
Write your markdown. Pick a template. Publish instantly.

Chart.js Charts
Chart.js renders data visualizations from JSON configuration inside chart fenced code blocks. Every LLM can generate Chart.js configs natively because it uses standard JSON syntax. Colors automatically adapt to your selected template.
Bar Chart
Compare values across categories with grouped bars. This example tracks monthly active users across two platforms.
Line Chart
Visualize trends over time with multiple series. The tension property controls curve smoothness.









Everything your markdown needs
Templates for the look. Destinations for the reach. Publishing for the world.
62 templates. Zero design work.
Pick a template and your markdown transforms into a polished, responsive page. Business reports, developer docs, academic papers, creative portfolios.
Browse templatesPaste-ready for every app.
Copy to Google Docs, Word, Slack, OneNote, Email, or Plain Text. Each destination gets purpose-built formatting that actually works when you paste.
See all destinationsPublish instantly. Share a URL.
One click to publish. Get a clean URL you can share anywhere. Custom slugs, analytics, and SEO included.
Learn about publishingCopy once. Paste perfectly.
Each destination gets its own formatting engine. What you paste actually looks right.
Project Update: Q1 2026
Last edited 2 minutes ago
The migration to the new API is 95% complete. All critical endpoints have been tested and validated against the staging environment.
Key Milestones
- Authentication service rewritten in Rust
- Database latency reduced from 120ms to 8ms
- Zero-downtime deployment pipeline live
| Task | Status | Owner |
|---|---|---|
| API migration | Complete | @chen |
| Load testing | In Progress | @park |
Link your thinking. See the connections.
Connect documents with [[wikilinks]], explore your knowledge graph, and let AI surface relationships you missed.
Wikilinks
Type [[ to link any document. Autocomplete finds it instantly. Rename propagates everywhere.
Learn about wikilinksDocument graph
See every connection in your workspace. Full-screen 2D and 3D views with folder-based clustering.
Explore the graphAI-powered linking
Related documents, unlinked mentions, and link suggestions. Powered by your own AI keys.
See AI featuresProfessional templates for every use case
62 professionally designed templates. Customize fonts, colors, and spacing. Or build your own.
Emergent Behavior in Multi-Agent Reinforcement Learning Systems
Department of Computer Science, Stanford University. Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy.
Abstract
We investigate emergent cooperative strategies in multi-agent reinforcement learning (MARL) environments with partial observability. Our framework, CoopNet, introduces a shared latent communication channel that enables agents to develop implicit coordination protocols without explicit reward shaping.
Note
All experiments were conducted on a cluster of 128 A100 GPUs over 14 days. Total compute cost: approximately $48,000.
Key Results
| Environment | Baseline | CoopNet | Improvement |
|---|---|---|---|
| Predator-Prey (10v10) | 62.4% | 89.1% | +26.7pp |
| Resource Gathering | 0.71 | 0.93 | +31% |
| Traffic Control | 847 avg wait | 312 avg wait | -63% |
Methodology
We extend the centralized training, decentralized execution (CTDE) paradigm with a variational communication bottleneck. During training, agents share a compressed latent vector z ~ q(z|o_i) of dimension d=64. During execution, each agent conditions its policy on both local observations and received messages.
The most sophisticated coordination we observed was not designed. It emerged from the pressure to succeed together.
Thesis
Academic
The Neon Renaissance
How digital artists are reclaiming the aesthetic of 1980s Tokyo, Miami Vice, and Blade Runner to create a visual language for the AI era. The glow is back, and it is intentional.
Color Theory in the Dark
Neon works because it exploits simultaneous contrast. A hot pink against deep navy creates a vibration that flat design cannot replicate. The eye is drawn to light sources instinctively.
Darkness is not the absence of design. It is the canvas that makes light meaningful.
Key Movements
| Movement | Signature | Origin |
|---|---|---|
| Synthwave | Grid horizons, chrome type | Music culture |
| Cyberpunk | Glitch effects, data overflow | Dystopian fiction |
| Vaporwave | Marble busts, palm trees | Internet irony |
| Neo-Tokyo | Dense vertical cityscapes | Anime, architecture |
Tip
When designing with neon, never use more than 3 glow colors. One primary, one accent, one highlight. More than that becomes noise.
The Palette
- Electric pink:
#FF2D9Bfor primary glow - Cyan:
#00F0FFfor secondary highlights - Deep navy:
#0A0E27for backgrounds - Purple haze:
#7B2FBEfor atmospheric gradients
Neon Nights
Creative
Getting Started with Markdown Publishing
Write in markdown, publish beautifully. Unmarkdown™ transforms raw markdown into polished, professional documents with one click.
Tip
Paste any AI-generated markdown and pick a template. Your document is ready to share instantly.
How It Works
- Write or paste your markdown content
- Choose a template from 62 professional designs
- Publish to a shareable URL, or copy to Google Docs, Word, Slack, and more
Supported Destinations
| Destination | Format | Preserves |
|---|---|---|
| Google Docs | Rich text | Headings, tables, images |
| Microsoft Word | Rich text | Full formatting |
| Slack | Slack markdown | Bold, code, links |
| Inline CSS | All visual styles |
The best formatting tool is the one that disappears into your workflow.
What You Can Format
- Headings with proper hierarchy and spacing
- Code blocks with syntax highlighting
- Tables with clean borders and alignment
- Task lists with interactive checkboxes
- Callouts for notes, warnings, and tips
Swiss
Free
Building a Real-Time Event Pipeline
This guide walks through designing a high-throughput event pipeline using Kafka, Flink, and ClickHouse. We process 2M events/second with sub-100ms latency.
Architecture Overview
Producer → Kafka → Flink → ClickHouse
↓ ↓
Dead Letter Grafana
Queue Dashboard
Warning
Always set replication: 3 for production topics. Single-replica topics risk data loss during broker failures.
Configuration
kafka:
brokers: ["kafka-01:9092", "kafka-02:9092"]
topic: events.raw
partitions: 32
replication: 3
flink:
parallelism: 16
checkpoint_interval: 30s
state_backend: rocksdb
Performance Benchmarks
| Component | Throughput | Latency (p99) |
|---|---|---|
| Kafka Ingestion | 2.1M/s | 12ms |
| Flink Processing | 1.8M/s | 45ms |
| ClickHouse Write | 1.6M/s | 89ms |
Dracula
Developer
Board of Directors: Q1 2026 Summary
This executive summary presents quarterly performance, strategic initiatives, and forward guidance for shareholder review.
Financial Performance
| Indicator | Target | Actual | Variance |
|---|---|---|---|
| Revenue | $14.2M | $15.8M | +11.3% |
| EBITDA | $3.1M | $3.4M | +9.7% |
| Customer Count | 2,400 | 2,680 | +11.7% |
| Net Retention | 115% | 122% | +7pp |
Tip
Net retention above 120% indicates strong product-market fit. Expansion revenue is outpacing churn by a healthy margin.
Strategic Priorities
- Enterprise expansion: Close 8 deals in pipeline worth $4.2M ARR
- Product differentiation: Launch AI features ahead of competitors
- International: Open London office by end of Q2
We are well positioned to accelerate growth while maintaining disciplined unit economics.
Risk Assessment
| Risk | Likelihood | Impact | Mitigation |
|---|---|---|---|
| Competitor pricing pressure | Medium | High | Accelerate feature roadmap |
| Key hire delays | High | Medium | Engage executive recruiters |
| FX headwinds (APAC) | Low | Low | Hedge quarterly |
Executive
Business

0
template variations
0
diagram and chart types
0
AI quick actions
0
paste destinations
0
template variations
0
diagram and chart types
0
AI quick actions
0
paste destinations
12 AI actions. One click each.
Fix grammar, adjust tone, add structure, translate. No prompting required.
Input
so basically the new auth system is way faster and we fixed alot of the bugs that were causing issues for ppl. the team did great work on this tbh and we should be good to ship it by friday
Output
The new authentication system delivers significantly faster performance. We resolved the key bugs that were impacting users. The team executed exceptionally well, and we are on track to ship by Friday.
Built for developers
13 REST endpoints. Full OpenAPI spec. llms.txt for AI tools. Publish markdown programmatically and integrate with your workflow.
13
REST endpoints
62
Templates
<200ms
Avg response
curl -X POST https://api.unmarkdown.com/v1/documents/publish \
-H "Authorization: Bearer um_your_api_key" \
-H "Content-Type: application/json" \
-d '{
"markdown": "# Hello World\n\nYour content here.",
"template": "minimal",
"slug": "hello-world"
}'Simple pricing
No credit card required. Upgrade when you're ready.
Free
For individuals getting started
- 5 documents, 3 published pages
- 8 free templates
- 10 AI actions (lifetime)
- All 6 paste destinations
- Mermaid diagrams + KaTeX math
- 3 shares per document
- 1,000 API calls/month
Pro
per month, billed annually
- Unlimited documents
- All 62 templates + custom editor
- Unlimited AI actions (30/hr)
- All diagram types (+ Graphviz, Chart.js)
- Access controls (public, link, specific people)
- Unlimited shares + folders
- Custom slugs + hide branding
- Downloads (DOCX, HTML, PDF)
- Full analytics + 1GB image storage
- 10,000 API calls/month

Your markdown deserves a beautiful home.
Start publishing for free. Upgrade when you need more.