Our team has gained momentum since winter and achieved high productivity consistently. Obviously, we should slow down our pace someday, but this April was not an exception.
We finished April with 100% uptime. We have surpassed over 900,000 registered devices on the platform and are now approaching the first million. It will not be a culmination but a remarkable achievement for us.
As promised in March’s changelog, we committed the initial release of geofences as an entity on the platform. Its UI is still hidden under the Experimental panel settings checkbox for most of our users, but you may explore the geofences API in the ApiBox already. The crucial thing about these geofences is that they are becoming a part of the flespi platform and are automatically integrated into all its systems. You will be able to:
- Enrich geofences with your own properties in metadata - specify color, on/off schedule, speed limits, priorities, common names, etc.
- Assign geofences to streams, plugins, calculators, and access and test device location inside all their expressions. For example, you will be able to apply plugins or streams only when the device is located in a geofence with some attributes and so on. Absolutely unlimited use cases.
- Use groups to automate geofences assignment and management.
- Access geofences REST API selectors. For example, in a single API call, retrieve devices that are located in or outside geofences and so on.
- Share geofences between multiple subaccounts using grants.
- Control access to each single geofence or by the group using configurable ACL in tokens.
Given the high volume of functionality provided, these geofences are available not for free. 1,000 geofences are included in all commercial plans, and each subsequent geofence over 1,000 will cost 0.01 EUR.
The previous geofence-related functionality, which allows you to define geofences as part of plugin or calculator configuration, will continue to operate as is. You may define as many geofences as you need, and we do not charge for them. We only limit and charge for these new geofences that are available as a dedicated entity on the platform.
Please be aware that this functionality is in experimental mode, and we may silently change the API until the official announcement in the Blog. Right now, there is no useful functionality available except the possibility to define the geofences via API and assign them to plugins, groups, streams, and calculators. But you cannot access these geofences in their expressions so far. We plan to finish the whole work with geofences and document it by the end of May.
Another remarkable achievement of April is the integration of RTMP protocol in our video telematics engine with which we finished all our initial promises for video functionality. Now Queclink dashcam devices support live video streaming functionality as well.
The next video telematics feature on our radar will be the support of live video streaming from flespi using the WebRTC protocol on par with currently available HLS. WebRTC protocol will enable live streaming into web browsers or mobile phones with minimal latency - 1-2 seconds maximum. This will be automatically available for all integrated video devices.
We released a plugin that allows our users to enhance device messages with geolocation services provided by Unwired Labs.
Also, WanWay devices have been integrated into flespi.
There was a large set of updates and new features for flespi analytics:
- Access to device metadata in expressions, which unlocks unlimited per-device configurations inside a single calculator.
- New functionality to enumerate all unique BLE beacons or visited geofences during the interval.
- New aggregation functionality for complex scenarios with multiple calculators involved.
We introduced a new tool in our flespi.io panel for a quick import of the devices list from a CSV file. And just in case you didn’t know, we documented how to export the result of a REST API call (e.g., list of devices, sorted by storage for example) to CSV.
We released Devices on Map widget for those who are using MQTT Tiles dashboard as a no-code visualization tool for data from flespi.
And last but not least, a few updates about codi, our sleepless multilingual AI assistant. By the way, in a very short 6 months time frame, we developed it from zero level up to a cool assistant that is now backing up to 40% of our support operations. And as everything in flespi, it is made fully in-house, except obviously LLM engine by itself.
We provided codi access to billing information, various account limits and counters, gave it the possibility to read entity logs when solving issues with an entity. It is really a joy now to read how it often correctly and deeply analyzes the problem and provides often very good recommendations and advice.
We trained codi to use the information from Teltonika Telematics Wiki for each of their device models and you can consult codi about very device-specific questions. Now if you are using a Teltonika device - codi knows almost everything about it - its current telemetry, settings, latest logs, configuration and even enriched with the knowledge from the official manuals from the manufacturer. Our next steps will be to load corresponding knowledge for other manufacturers and obviously we will further train codi on how to combine all these knowledge and tools together to provide really deep device analysis and generate cool recommendations.
And finally, in less than a couple of months from our AI hackathon, we released a PVM generator tool that you may use to convert natural or programming language instructions into PVM code which can transform device messages according to your algorithms.
In May, we will complete the integration of new geofences into various systems, add additional functions to the analytics engine, continue with realms enhancement, and will continue working on two of our strategic areas for 2024 - AI and video telematics. Stay tuned!