Make AI Agents
Make AI Agents app
22 min
make ai agents is available on all plans using make's ai provider , with the option to use custom ai provider connections on paid plans make ai agents is available in open beta as a beta feature, both product functionality and pricing may change t he make ai agents app, in the {{scenario singular}} builder, allows you to use ai agents in {{scenario plural lowercase}} this article outlines the app's modules and their credit usage modules run an agent the make ai agents > run an agent module allows you to create ai agents, provide agents with tools, and send messages to agents — either as additional instructions or tasks the following outlines the fields in the m odule settings , found by clicking the module agent in agent , c hoose an existing agent or click the create agent button to create a new agent in the connection field, select your connection in the agent name field, fill in your agent's identification in the model field, select the language model and version provided by your ai service provider in the system prompt field, define the agent's purpose and constraints the system prompt defines the agent in all the {{scenario plural lowercase}} it uses in each {{scenario singular lowercase}} , you can further customize the agent's description with additional instructions context in context , you can view the context files the agent uses for reference thread id in thread id , you can specify a custom thread id to keep all of your interactions with the agent in the same communication thread if you leave the thread id field empty, a new thread id appears in the output the agent will have no memory of previous communication examples of what to map to the thread id field the userid to remember conversations with a user the timestamp of the first message of a continuing thread to remember the thread the more context an agent must remember in the thread id , the higher the tokens in the agent settings of an agent configuration, you can limit the number of agent runs in the thread history to reduce context from the thread id additional system instructions in additional system instructions , you can share extra contextual information with the agent dynamically provide context by mapping data from the {{scenario singular lowercase}} to this field messages in messages , you can s end tasks and specific requirements to the agent you can send multiple user messages at once timeout in timeout , you can set the maximum number of seconds the agent can execute until it is terminated if you leave the timeout field empty, the timeout is 300 seconds (5 minutes) the maximum timeout is 600 seconds (10 minutes) output schema in output schema , you can define the output schema of the agent response text the response is text make schema the response is based on the output structure you define if you click generate , {{product name}} generates a structure based on a content type and sample data the following outlines the fields in the module output , found by clicking the output bubble after running the module response in response , you can view the agent's response to the user request optionally, mapping docid\ upfwozscc5soixr2954z2 this data to the next modules executionsteps in executionsteps , you can expand each step to view the agent's thinking and decision making process steps are in chronological order and describe any of the following factors role role providing the reasoning step, which includes the system (prompt), user , assistant (ai agent), or tool content information processed at a given step of the agent's reasoning, for example, a text input reasoning summary of the agent's thought process before responding to the user, for example "i retrieved the value of 'alpha,' which is 111 now, i need to multiply it by 3, giving me a new value of 333 for 'beta ' i'll set this using the function functions set variable 1, specifying 'beta' as the module name and '333' as the module value after that, i want to inform the user concisely about the value saved i'll go ahead and call the function now!" agent iterationid the id shared by all of the agent's steps towards producing an output toolcalls tool called, such as a scenario or module tool, and its identifying information tokenusage ai tokens used in a single module run, including prompttokens (input), completiontokens (output), and total tokens toolresponse response of the tool called, including the information retrieved and execution time tokenusagesummary in tokenusagesummary , you can view ai tokens used in a single module run, including prompttokens the input tokens processed, such as for prompts, tool responses, tool descriptions, and arguments completiontokens the output tokens processed, such as for content, reasoning, and tool calls totaltokens the tokens processed overall create agent context (file) the make ai agents > create agent context (file) module allows you to upload a context file to an ai agent this module is an alternative to uploading context files in the context field of the ai agents configuration tab to use create agent context (file) in a {{scenario singular lowercase}} in the scenario builder, add a module that retrieves the context file, for example, a google drive > download a file module add the create agent context (file) module in the create agent context (file) module's configuration, under agent , select an existing agent that will store the file as context alternatively, create a new agent in file , select the preceding module that retrieved the file, or map specific data from the file in map click save click run once you have now uploaded a context file to your ai agent you can view the file under the context field of the ai agents configuration tab to replace an uploaded context file with an updated version that has the same file name, use an upsert agent context module agent select an existing agent that will store the context file, or click the create agent button to create a new agent context upload a context file that your selected agent can access you can also see the files you already uploaded to the agent file choose the name of the module that downloaded the context file this option uploads the entire file and keeps the same file name as the original map change the file name in file name and map a specific value from the context file in data create agent context (text) the make ai agents > create agent context (text) module allows you to upload text to an ai agent to be used as context agent select an existing agent that will store the context file, or click the create agent button to create a new agent context name name the context file to identify and retrieve it later context insert text or map text from preceding modules get agent contexts the make ai agents > get agent contexts module retrieves all the context files uploaded to an ai agent when you run the module, context files appear in the module's output to view its output, click on the white bubble above the module under operation , expand body , which includes a numbered item for each context file agent select an existing agent, or click the create agent button to create a new agent delete specific agent context the make ai agents > delete specific agent context module allows you to delete a specific context file from an ai agent's memory agent select an existing agent, or click the create agent button to create a new agent context select an uploaded context file to delete from an agent's memory delete all agent context the make ai agents > delete all agent context module allows you to delete all context files from an ai agent's memory you can use this module on its own in a scenario click run once to delete all of the agent's context files agent select an existing agent, or click the create agent button to create a new agent upsert agent context (file) the make ai agents > upsert agent context (file) module allows you to upload or replace a context file this module is ideal when you need to replace an existing file with a file of the same name to use upsert agent context (file) in a {{scenario singular lowercase}} in the scenario builder, add a module that retrieves the context file, for example, a google drive > download a file module add the upsert agent context(file) module configure the module settings and click save click run once you have now upserted a context file you can check the file in the ai agents configuration tab, in context agent select an existing agent that will store the context file, or click the create agent button to create a new agent context upload a context file that your agent can access you can also see the files you already uploaded to the agent file choose the name of the module that downloaded the context file this option uploads the entire file and keeps the same file name as the original map change the file name in file name and map a specific value from the context file in data upsert agent context (text) the make ai agents > upsert agent context (text) module allows you to insert updated text to replace an existing text context this module is ideal when you need to replace existing text that has the same name to use upsert agent context (text) in a {{scenario singular lowercase}} , see the steps outlined in the upsert agent context (file) section agent select an existing agent that will store the context file, or click the create agent button to create a new agent context name name the context file to identify and retrieve it later context insert text or map text from preceding modules credit usage the make ai agents app uses credits docid\ sqbfn9s1f2nbq71kqhq7x differently depending on the module and whether it uses credits docid\ hh nrtakeqakbmg63uuin or a custom ai provider connection (e g , via an api key) users on all plans can use make's ai provider users on paid plans can also use custom ai provider connections , such as openai and anthropic claude the table below outlines make ai agents modules and their credit usage logic module uses make's ai provider (all plans) uses a custom ai provider connection (paid plans) run an agent 1 credit per operation + credits based on ai token usage 1 credit per operation create agent context (file) 1 credit per operation + credits based on token usage, including ai tokens (file description generation) embedding tokens (vectorization) 1 credit per operation + embedding tokens (vectorization) create agent context (text) 1 credit per operation + credits based on token usage, including ai tokens (file description generation) embedding tokens (vectorization) 1 credit per operation + embedding tokens (vectorization) get agent contexts 1 credit per operation 1 credit per operation delete specific agent context 1 credit per operation 1 credit per operation delete all agent context 1 credit per operation 1 credit per operation upsert agent context (file) 1 credit per operation + credits based on token usage, including ai tokens (file description generation) embedding tokens (vectorization) 1 credit per operation + embedding tokens (vectorization) upsert agent context (text) 1 credit per operation + credits based on token usage, including ai tokens (file description generation) embedding tokens (vectorization) 1 credit per operation + embedding tokens (vectorization) note the following specifics about create agent context and upsert agent context modules embedding tokens are used when {{product name}} converts chunks of a context file into numbers (vectors) to be stored in make's rag database, turning the file into a retrievable format for ai agents the number of embedding tokens will depend on the file size ai tokens for file description generation are used when an ai provider summarizes a context file based on the first part of the file when you use make's ai provider, {{product name}} bills you for these ai tokens if you use a custom ai provider connection, the provider bills you for the tokens when you trigger an agent in a scenario through testing & training in the ai agents configuration tab, this action uses 1 credit for each chat input, in addition to each operation (and tokens, when relevant) used