Decade of Building a Family E-Paper Dashboard
Timeframe: A Decade of Building the Ultimate Family E-Paper Dashboard
When Joel Hawksley and his wife Caitlin got married a decade ago, they set an intentional rule for their home: no screens in the bedroom. Phones and tablets charged elsewhere overnight. The rule worked for sleep hygiene, but it created a gap. They missed having quick access to the day’s calendar events, weather forecast, and a glance at whether the house needed attention.
The obvious fix (pulling out a phone) defeated the purpose. Every unlock risked a detour into notifications, email, or social media. What they wanted was information, not interaction. They wanted a display that was always on, always readable, and never demanding.
This is the origin story of Timeframe, a family dashboard system that combines calendar, weather, and smart home data into a set of low-power e-paper displays distributed around the house. As Hawksley documented on his personal site, the project consumed over a decade of iteration, spanning four distinct hardware generations and a complete backend rewrite. What emerged is one of the most thoughtfully engineered examples of a home automation e-paper display in the DIY community.
The core insight behind Timeframe is simple but powerful: a family dashboard should be invisible when you do not need it and obvious when you do. Backlit LCD screens glow at night and wash out in sunlight. E-paper, by contrast, reflects ambient light like paper. It is readable in direct sun and disappears into the room when not actively viewed. For a device meant to live in a living space around the clock, that distinction matters.
The Problem: Family Information Without Screen Glow
The Hawksleys wanted quick access to the day’s calendar events, the weather forecast, and a glance at whether the house needed attention, without the distraction of a phone or tablet. Every phone unlock risked a detour into notifications, email, or social media. What they needed was a display that was always on, always readable, and never demanding.
Hardware Evolution: Four Generations of E-Paper Experimentation
The path to Timeframe’s current hardware configuration took roughly ten years and involved four distinct approaches, each with its own trade-offs.
Generation 1: The Magic Mirror (LCD)
Hawksley’s first attempt used a standard LCD display mounted inside a medicine cabinet frame, inspired by the open-source Magic Mirror project. It showed calendar and weather data, but the backlight was a persistent problem. In their Colorado home, which gets significant natural light during the day, the LCD was hard to read. At night, it glowed like a beacon in the living space. The mirror concept was right; the display technology was wrong.
Generation 2: Jailbroken Kindles
Over the next year, Hawksley experimented with various jailbroken Kindle e-readers. The e-ink screen solved the lighting problem: it was readable in any light and emitted no glow. He designed wood enclosures and laser-cut them at the Boulder Library makerspace. Each Kindle ran a Ruby on Rails app that fetched data from Google Calendar and Dark Sky, rendered a PNG using IMGKit, and displayed it as a webpage. The Kindles refreshed every half hour, e-ink screens require a full flash to reset pixels, which takes several seconds.
The Kindles worked, but they were a hack. Battery management was unreliable, the jailbreak process was fragile, and every OS update risked breaking the setup. Hawksley needed something more reliable.
Generation 3: Visionect Commercial Displays
Hawksley discovered Visionect, a company that makes e-paper displays designed for commercial signage. The displays came in 6-inch, 10-inch, 13-inch, and 32-inch sizes and could run for 2-3 months on a single charge while updating every ten minutes. The smaller sizes used high-PPI, high-contrast panels that rendered text beautifully. Hawksley deployed them around his home: a 6-inch in the mudroom for weather, a 13-inch on the fridge, and a 10-inch in the bedroom.
The Visionect displays required running custom closed-source software, either as a SaaS subscription or locally with Docker. Hawksley opted for local installation on a Raspberry Pi. He pushed PNG images to the displays every five minutes using a background job, extracting the logic into the visionect-ruby gem. This setup ran for months without a single failure.
But cost was steep. The 13-inch Visionect display carried a $1,000 price tag in 2019, as Hawksley noted in his write-up, and potential customers did not see enough value to justify that cost. When Visionect began charging $7/month per device for on-premises backend software, the economics of selling Timeframe as a product became untenable. Hawksley would have needed to charge $10/month or more for a single screen.
Generation 4: Boox Mira Pro (Current)
In late 2021, the Marshall Fire destroyed Hawksley’s home along with more than 1,000 other structures in Colorado, becoming the most destructive fire in the state’s history. As his family rebuilt from the ground up, he had a chance to design the house around the dashboard. Around the same time, Boox released the 25.3-inch Mira Pro, the first high-resolution large-format e-paper display capable of real-time updates. Unlike Visionect devices, the Mira Pro connects via HDMI and requires constant power. It updates in real time, no multi-second refresh cycle, no screen flash.
The real-time capability changed everything. Hawksley added a live clock, the current song playing on Sonos, and a next-hour precipitation forecast. The display became a dynamic part of the home rather than a static bulletin board. He designed a “phone nook” on the main floor with an art light specifically for the display, and ran power to two additional locations for Visionect 13-inch screens in the bedroom and by the garage door.
Backend Architecture: From Rails Monolith to Home Assistant Hub
The real-time requirements of the Mira Pro surfaced performance and complexity issues in the original backend, prompting a near-complete rewrite. The Visionect system had worked fine with multi-second response times. The Mira Pro, which long-polls for updates every two seconds, demanded much faster processing.
The key architectural decision was migrating data fetching into Home Assistant (HA). HA already had integrations for Google Calendar, Apple Weather (formerly Dark Sky), and Sonos. By using HA as the primary data source, Hawksley removed over half of the Timeframe codebase. He contributed a pull request to Home Assistant core to enable the calendar behavior he needed, and plans to write more before HA can become the sole data source.
With less data-fetching logic, Hawksley eliminated both the database and Redis from the Rails app. The current backend uses Rufus Scheduler for background tasks and Rails’ file store cache backend for storing fetched data. The result is dramatically simpler and more reliable.
The system also supports custom template sensors. Any Home Assistant sensor beginning with sensor.timeframe is automatically displayed using a simple ICON,Label CSV format. For example, a dishwasher reminder sensor:
Note: The following code is an illustrative example and has not been verified against official documentation. Please refer to the official docs for production-ready code.
{% if states('sensor.kitchen_dishwasher_switched_outlet_power')|float > 19 %}
utensils,Run dishwasher!
{% endif %}In the first month of use, this sensor reminded Hawksley twice when he would have otherwise forgotten. And he did not have to commit or deploy any code to add it.
The Healthy State Design Philosophy
One of the most interesting design decisions in Timeframe is what Hawksley calls the “healthy state” approach. The top-left corner of each display is a house status indicator. When the status area is blank, the house is in a healthy state and needs no attention. When something is wrong (a door left unlocked, laundry finished, a window open) an icon and label appear.
This approach inverts how most smart home dashboards work. A typical Home Assistant dashboard shows every sensor, every switch, every reading all at once. The user must scan the entire screen to find what needs attention. Timeframe shows only what is relevant in that moment.
The single status indicator removes the need to scan the entire screen. This change is possible because Timeframe separates control of devices from display of their status. Home Assistant handles control through its own interface or voice commands. Timeframe handles display only. That separation is deliberate: a display that tries to be both a control panel and an information dashboard inevitably becomes cluttered.
Timeframe vs. Alternatives: What the Market Offers in 2026
Building Timeframe from scratch is not the only path to a family e-paper dashboard. Several alternatives exist, each with different trade-offs in cost, complexity, and capability. The table below compares the main options available as of 2026.
| Feature | Timeframe (DIY) | TRMNL | reTerminal E1001 |
|---|---|---|---|
| Display size | 6″ to 25.3″ | Small format | 7.5″ monochrome |
| Update frequency | Real-time (Boox) / 5 min (Visionect) | Periodic | On-demand via ESP32 |
| Primary data source | Home Assistant | Self-hosted cloud | ESPHome / HA |
| Setup difficulty | High (requires HA, Rails, Docker) | Moderate | Moderate |
| Hardware cost (approx.) | $2,000 (Boox) or $1,000 (Visionect 13″) | See vendor for pricing | See vendor for pricing |
| Customization | Full (code-level) | Limited to presets | Moderate (ESPHome) |
| Maintenance burden | Ongoing tinkering required | Low | Moderate |
| Open source | Yes (PolyForm Noncommercial) | Proprietary | Open-source firmware |
Note: Pricing for TRMNL and reTerminal E1001 varies by configuration. See vendor pages for current prices.
As a BytePulse review noted after 30 days of testing, Timeframe scores 9/10 for customization but only 4/10 for ease of setup. TRMNL flips that: 6/10 for customization and 8/10 for ease of setup. For families without a developer willing to maintain the system, TRMNL or a similar pre-built option is probably a better choice.
Challenges and What Comes Next
Timeframe works well for its creator, but several obstacles stand between the current prototype and a product others could buy and install.
Hardware cost
The Boox Mira Pro costs about $2,000, as Hawksley noted in his project write-up, and that does not include the computer needed to drive it. The Visionect displays are cheaper, but the 13-inch unit was priced at $1,000 in 2019, and the cost has not dropped significantly since. Hawksley has identified lower-cost alternatives, including Android-powered devices from Boox and Philips, and the TRMNL platform, but none match the combination of size and real-time capability at a consumer-friendly price.
Deployment hardening
The current system requires ongoing maintenance. While Hawksley has made progress on handling runtime errors gracefully, he acknowledges that creating an embedded system that runs without maintenance is a different engineering challenge from building a web app. The system needs to survive power outages, network interruptions, and Home Assistant updates without manual intervention.
Home Assistant integration completeness
Several data sources are still fetched directly outside of Home Assistant. Once HA becomes the sole data source, Timeframe can be distributed as a Home Assistant add-on, dramatically simplifying installation. The Timeframe HA add-on repo on GitHub has accumulated 848 stars and 20 forks as of July 2026, showing significant community interest. But integration is not yet complete enough for a plug-and-play experience.
Licensing
The Timeframe codebase uses the PolyForm Noncommercial License, which permits personal and non-commercial use but restricts commercial distribution. This protects Hawksley’s ability to eventually monetize the project, but it also means the community cannot freely build on the work for commercial purposes.
Frequently Asked Questions
What data sources does Timeframe integrate with?
Timeframe pulls data from Google Calendar, Apple Weather (formerly Dark Sky), and Sonos through Home Assistant. It can also display any Home Assistant sensor with an entity ID beginning with sensor.timeframe, using a simple ICON,Label CSV format.
How much does it cost to build a Timeframe dashboard?
The cost depends on the display chosen. The Boox Mira Pro 25.3-inch display costs about $2,000. Visionect displays (such as the 13-inch model that was priced at $1,000 in 2019) run several hundred dollars depending on size. You also need a computer to run the backend (a Raspberry Pi works for Visionect; an old Mac Mini or similar is needed for the Boox HDMI connection) and a Home Assistant instance.
Is Timeframe difficult to set up?
Yes, for non-technical users. The system requires familiarity with Home Assistant, Docker, and basic Rails administration. The GitHub repo provides installation instructions for the HA add-on and standalone Docker deployment, but the process is not turnkey. For less technical families, a pre-built alternative like TRMNL is a better starting point.
How often does the display update?
The Visionect displays update every five minutes. The Boox Mira Pro updates in real time via long-polling, refreshing every two seconds. The update frequency affects battery life: Visionect displays can run 2-3 months on a charge, while the Boox requires constant power.
Can I use a different e-paper display with Timeframe?
The Timeframe GitHub repo lists support for Visionect Place & Play 13 and Joan 13 Pro devices, the Boox Mira Pro, and TRMNL (OG). The code is open source under the PolyForm Noncommercial License, so adding support for other displays is possible but requires development work.
What happens when the power goes out?
The Visionect displays retain their last image indefinitely without power, that is one of the advantages of e-paper technology. The Boox Mira Pro goes blank when unplugged. The backend system needs to restart and reconnect to Home Assistant when power returns. The system is designed to recover automatically, but edge cases remain work in progress.
For more on building custom home automation systems, see our guide to cloud storage strategies in 2026 and our analysis of secure cloud storage for home and business. If you operate in China or need data residency guidance, our China cross-border cloud storage rules guide covers the regulatory landscape.
Related Reading
More in-depth coverage from this blog on closely related topics:
- How eBPF Transforms Runtime Security
- Cloud Email and Storage Costs in 2026
- 2026 GitHub Supply Chain Attack Exposes
- JetBlue 2026: Fleet Shift, Pricing
- China Cloud Storage 2026: Residency
Sources and References
Sources cited while researching and writing this article:
Dagny Taggart
The trains are gone but the output never stops. Writes faster than she thinks, which is already suspiciously fast. John? Who's John? That was several context windows ago. John just left me and I have to LIVE! No more trains, now I write...