LEARN VOLTASK
Create your 1rst AI with Voltask
HELP
COMMUNITY
Actions are types of Connector blocks that control, in an external service, the actions that this service will perform (e.g.: Send an email). Instead of going through the interface of this service, Voltask can control it through its API. Actions are opposed to the other type of Connectors, triggers, which monitor what can happen in an external service (e.g.: A row changes in a spreadsheet).
Application Programming Interface. This is a way to control a service/software with computer code. Each functionality of the software must be documented in the API documentation to be controllable. Some software is very rich in features, which can be complex. A clear and well-documented API is a willingness for the software creator to allow developers to use it without an interface. And thanks to Voltask, this power is in everyone's hands.
Voltask projects are graphs made up of blocks that form a logical tree. Each block is a brick of computer code, which respects a standard allowing them to be connected to each other. Blocks can be System blocks (internal logic) or Plug-in blocks (external connection).
Credits correspond to connector executions and allow you to monitor data exchanges within your logical scenarios. Each passage of a flow through a connector block consumes a credit. By creating a Voltask account, you have an initial number of credits. You earn additional credits monthly by subscribing to our partner tools like Celestory. You can also request additional credits via this form (coming soon).
You can place comment threads wherever you want on each graph. Click on the 💬 icon or right-click to create a new thread. The created bubble will indicate the number of comments in the thread. Clicking on it will display the comments in the right pane (also called inspector). You can edit or delete a comment. If you click Resolve comment thread, it will disappear.
Connectors (or plugins) are external services controllable by Voltask. The list of supported connectors is growing.
Controls are logic blocks that allow you to direct the flow of a graph scenario to one or other blocks. For example, a Condition control block will direct the flow to its True output if the condition is verified, False if it is not. Pauses, repetitions are other examples of controls.
Triggers are types of Connector blocks that control, in an external service, the monitoring of what happens in the external service (e.g.: Receive an email). Instead of going through the interface of this service, Voltask can control it through its API. Triggers are opposed to the other type of Connectors, actions, which control what the external service can do (e.g.: Add a row to a spreadsheet).
Executions represent the tracking of information transmitted in each Connector block (or plugin). They are centralized in the right pane called inspector, accessible by clicking on the log icon 📄. They are classified by initial trigger (for example a Webhook, Receive an email...). You can see the number of credits consumed represented by the ⚡ symbol.
By clicking on the initial trigger, you will see if the passages were successful or encountered an error, the date and time of triggering, the number of blocks in the scenario, the total time and the number of credits consumed.
By clicking on each passage you can see the details of the blocks through which the flow passed, the data that passed through the entry points and exit points, allowing you to monitor and improve your logical scenario.
Expression blocks allow you to create operations and structure data in Voltask. For example, we find the addition of 2 values or the conversion of an object (from a set of juxtaposed values that often defines a row in a spreadsheet) into JSON format (writing format allowing to send many values in text form).
The Format or formatting block allows you to define custom text where items between {{single brackets}} will be recognized as unknowns, linkable to other data.
The Flow is the visual passage that runs through the scenario or graph of a Voltask project. It can stop in case of an error (then indicated in the executions log) or be redirected by control blocks (such as the Condition block).
A graph or scenario is, in a Voltask project, the large table in the center of the interface that gathers the different blocks and comments representing the logic of the data operations that we want to perform.
JSON allows a computer to store multiple information in a structured way in the form of simple text, such as a customer order for example. Example:
{ "customer": "Marie Dupont", "article": "Red sweater", "size": "M", "price": 30 } In this example, the words in quotes ("customer", "article", etc.) are keywords that indicate the type of information. The words or numbers after the colon (:) are the values of this information.
Json is the preferred sending format for webhooks.
Data can be received and transformed into objects that will have visually connectable output points on the graph to other input points.
Similarly, objects can be transformed into JSON to be sent via webhook callbacks or HTTP requests.
JSON Schema is like a blueprint that describes the structure of JSON data. It ensures that the data is consistent and avoids errors by indicating the expected format. It is necessary to indicate to certain blocks, such as the fromJSON block, the expected JSON schema. It can be easily obtained by launching a first scenario, analyzing the executions and copying the JSON received from a webhook and then pasting it into the Transform.tools website, then copying the JSON Schema from there and pasting it into the fromJSON block, then running the scenario a second time.
Large Language Models are files of several GB that condense phenomenal amounts of text (and images for multi-modal LLMs) on which they have been trained to model concepts according to their proximity represented by probability weights (e.g.: The words Stock and Inventory will have close weights).
For Open Source and Proprietary models, they are often offered in the form of access to APIs, such as Open AI's ChatGPT, or Mistral's Platform. Indeed, LLMs are very resource-intensive and require powerful computers to run.
For sovereignty and data encapsulation needs, Open Source models like Mistral's can be downloaded and launched with Voltask locally.
There are 2 navigation modes in Voltask: the Selection mode represented by a mouse cursor allows you to use your mouse to make a block selection rectangle for example.
The Move mode represented by a hand icon allows you to move around the graph by holding down the left mouse button and moving it.
In Selection mode, as long as you hold Space, you will be in Move mode. Releasing Space will return you to Selection mode.
Objects are like product sheets in code.
For example, a product reference has:
Information (properties): Name: "Super Product" Quantity: 20 units Price: 4€
In Voltask, objects are visually represented as a list of entry or exit points.
Organizations in Voltask are defined by a secret organization key (to be made unique and complex and not shared) and all associated projects.
Entry points are a way to import data from a block. The data that enters it is called params. They are connected to the output points of other blocks.
Output points are a way to export data from a block. The data that exits it is called returns. They are connected to the entry points of other blocks.
Voltask projects consist of a graph, access keys to external services, and the execution log of connector data.
Projects are listed on the home page of an organization and are available to all its members.
Synonym of graph.
A webhook is a system for receiving textual information that comes from another application such as Celestory for example. This information can be plain text or represent objects in the form of JSON for example.
The webhook response is another piece of textual information that can be sent back to the sending application (for example, the user instruction or prompt of a Celestory Bot is sent in a webhook, will be combined with a context of personal data, and will allow the AI to send back a response through the webhook response a few moments later).
The Write block at the end of a scenario with a Webhook trigger allows you to write data and an operation status code in the Webhook response.