Relevance AI Tool Builder Transformations as Helper Functions
All transformations available in our tool builder are also supported from the our python code step.
Use a large language model such as GPT
Arguments
prompt
: str
The prompt that is fed to the model.model
: str
The model to use for completion. If using gpt3.5, if you do not set your own API Key you will be charged credits for usage. For other models, Make sure to set an API key.fallback_model
: str
The model to use for completion if the model selected in the main window fails for whatever reason. If using gpt3.5, if you do not set your own API Key you will be charged credits for usage. For other models, Make sure to set an API key.history
: Any
Conversation history to be passed into the prompt. For example, [{role: ‘user’, message: ‘Hello, my name is Bob.’}, {role: ‘ai’, message: ‘Hello Bob, how are you?’}].history_only
: bool
If true, prompt isn’t included in what gets sent to the llm. mostly useful for re-prompting an llm after an function callsystem_prompt
: str
System prompt to be passed into the GPT chat completion prompts.strip_linebreaks
: bool
Whether to strip linebreaks from the output.temperature
: int | float
Temperature of the selected model. Typically, higher temperature means more random output.response_format
: str
For certain models such as OpenAI, response format can be forced through this field.max_tokens
: int | float
How many tokens would you like the output to be?seed
: int | float
Seed for generated output. The seed value stores the nature of randomly generated output, the same seed with generated the same response regardless of other parametersvalidators
: Any
Validate that the LLM produces output in the expected format, and re-prompts to fix issues if not. Note: the LLM still outputs text, this only validates the shape of that text.split_and_map
: bool
If a prompt variable is too long, split the variable into chunks, and run the entire prompt on each chunk. Very helpful for processing large pdfs.memory
: Any
Define how variables should be handled if they’re too long. Choose from extracting the most relevant part, a summary of it or whether to use it all.functions
: Anyfunction_to_force_gpt_to_call
: str
The name of the function to force GPT to call. This is useful if you want force GPT to call a specific function.timeout_ms
: int | float
Timeout in milliseconds, request will timeout after this much time if it runs for too long.debug
: boolmock_llm_response
: dicttoken_logging
: dict
When provided, tokens will be logged based on these properties, and aggregatable by agent and use case.stream
: bool
Streams the output as it is generated. Only implemented when running an agent.azure_model
: str
The Azure OpenAI Deployment Id to use when picking the azure-openai custom deployment model.Returns
answer
: strprompt
: strvalidation_history
: Any
Conversation history to be passed into the prompt. For example, [{role: ‘user’, message: ‘Hello, my name is Bob.’}, {role: ‘ai’, message: ‘Hello Bob, how are you?’}].split_answers
: Anyuser_key_used
: boolcredits_cost
: int | floaterrors
: Anydebug
: dicthistory_items
: Any
Conversation history to be passed into the prompt. For example, [{role: ‘user’, message: ‘Hello, my name is Bob.’}, {role: ‘ai’, message: ‘Hello Bob, how are you?’}].model_used
: strRun an API request
Arguments
method
: str
The HTTP method to use.url
: str
The URL to make the request to.headers
: dict
The headers to send with the request.body
: Any
The body to send with the request.url_params
: dict
Enter any key-value pairs to append to the base URL as query parameters. It’ll be formatted like key1=value1&key2=value2. Useful for specifying attributes of data you’re requesting.response_type
: str
The format of the response.cookies
: dict
The cookies to send with the request.Returns
response_body
: Anystatus
: int | floatbody
: strurl
: strresponse_headers
: dictRun JavaScript code
Arguments
code
: str
The JS code to execute, returning the value to be used as the output of this transformation.backend
: str
Pick the runtime to use for this code.Returns
transformed
: Anyduration
: int | floatRetrieve data that best match a provided query
Arguments
dataset_id
: str
The dataset to use as knowledge. If the list is empty, please upload a dataset, and check to vectorize it when prompted.query
: str
Items that best match the query will be retrieved from the knowledge source.query_type
: str
The type of search to perform. Keyword search is better at finding names, and proper nouns, while vector search is better at finding synonyms and related ideas.vector_field
: str
The name of the field that contains the vector. Only applicable for vector search.model
: str
The model name to use.content_field
: stroutput_all_fields
: bool
If set to true, will always output all fields. If set to false, this setting will have no behaviour.page_size
: int | float
The number of results to return.raw_filters
: Any
Raw relevance filters to filter results searched on.Returns
results
: Anydocuments
: dictVectorise an array of strings and do similarity search
Arguments
array
: Any
The array of data to search. If it is an object, it will be stringified when searching.query
: str
The query to search for.page_size
: int | float
The number of results to return.field
: str
The field to search in if array includes objects.Returns
results
: Anyscores
: AnyInsert data into a knowledge set.
Arguments
dataset_id
: str
The knowledge set to insert data to.documents
: Anyclean_field_names
: bool
If set to true, field names will be cleaned before inserting. This is recommended to avoid issues with special characters in field names.ingest_in_background
: bool
If set to true, the step will immediately continue, without waiting for documents to be inserted and available.Returns
inserted
: int | floatfailed_documents
: AnyRun a custom Apify Job, and optionally save results to Knowledge
Arguments
input
: dict
The input object for the Apify Jobactor
: str
The actor to use, for example apify/web-scraperknowledge_set
: str
If provided, write results to a knowledge setoutput_settings
: dict
Controls the output options for the apify job, based on https://docs.apify.com/api/client/js/reference/interface/DatasetClientListItemOptionsdont_use_personal_key
: boolcost_target
: int | float
This is a cost target for the transformation to aim for.
We use a rough linear approximation to estimate how much the step will cost.
The true cost may be slightly over or under depending how noisy the sampling method is.
Does not apply to those who use their own Apify key.Returns
credits_cost
: int | floatitems
: AnyArguments
urls
: Any
URLs of Companies on Glassdoorknowledge_set
: str
Did you want to insert the results of the transformation into a knowledge set?search_keyword
: str
Search Keywordmax_items
: int | float
Number of results per Glassdoor search pageend_page
: int | float
Number of search pagesscrape_salary
: bool
Do you want to scrape salary information?custom_map_function
: str
A custom map function for scraping each page.extended_output_function
: str
A custom extended output function for scraping each page.Returns
documents
: Anycredits_cost
: int | floatArguments
urls
: Any
URLs of Companies on Glassdoorknowledge_set
: str
Did you want to insert the results of the transformation into a knowledge set?should_chunk_text
: bool
Should the text fields be chunked?Returns
status
: strdocuments_url
: strGenerate a vector embedding from a given input with a choice of models.
Arguments
input
: str
The input to generate a vector embedding with.model
: str
The model name to use.Returns
vector
: AnySend an email with subject and body
Arguments
recipientEmails
: Any
The recipient emailssubject
: str
What is the subject of the email you would like to send?body
: Any
What is the body of the email?Returns
result
: strstatus_code
: int | floatCombine an array of strings into a single string
Arguments
array
: Any
The array of strings to join.sep
: str
The separator to use between each element when joining the array.Returns
text
: strExtract text from PDF without OCR. Doesn’t support handwritting and scans.
Arguments
pdf_url
: str
The URL of the PDF to convert to text.use_ocr
: bool
whether to use Microsofts ocr recognition model to extract text from the page. If no custom azure key and domain are provided in project keys, it will charge 100 credits per page.ocr_type
: strreturn_as_documents
: bool
Toggling this will either return the PDF text as an aggregated text or as a list of texts.Returns
text
: strnumber_of_pages
: int | floatdocuments
: AnyRetrieve knowledge given a query and source.
Arguments
query
: str
Enter your question here. Tip: keep it terse to avoid ambiguity for clearer answers.top_k
: Any
No. of results to retrieve.knowledge_set
: str
Select the knowledge source. If empty, will run retrieval on your entire knowledge base.text_field
: Any
Select the text fields/columns to vectorize and run retrieval on. It will vectorize all fields, if unspecified.use_files
: bool
Select this if you want to run retrieval on raw files instead of pre-extracted text. Note: Set a knowledge set when using this.custom_prompt
: str
Enter a custom prompt to generate the response, or use the default prompt below.temperature
: int | float
The higher the temperature, the more random the output. Lower temperatures give more predictable results.llm_model
: str
The LLM model to use for generation.embed_provider
: str
The embedding model to use for retrieval.reranker_model
: str
The reranking model to use post-retrieval.chunking_strategy
: str
The strategy to use for chunking the input text.retrieval_strategy
: str
The strategy to use for retrieval. Use hybrid for best results.Returns
response
: strsources
: Anycitations
: Anyduration
: int | floatExtract text from an audio file using OpenAI.
Arguments
audio_url
: str
URL of the audio to transcribe.response_type
: str
Whether to return the transcript only, or also advanced metadata such as timestamps and confidence scores. If returning advanced metadata, the keys ‘language’, ‘duration’ and ‘segments’ can be accessed on the output object.Returns
text
: strlanguage
: strduration
: int | floatsegments
: AnyTranscribe an audio/video file, including speakers, timestamps and automatic grammar correction, with Deepgram.
Arguments
audio_url
: str
URL of the audio to transcribe.diarize
: bool
Identify when different people speak. Each word in the transcript will be assigned a speaker ID, starting from 0.Returns
metadata
: dictresults
: dictTranscribe an audio/video file, including speakers, timestamps and automatic grammar correction, with AssemblyAI.
Arguments
audio_url
: str
URL of the audio to transcribe.Returns
id
: str
The unique identifier for the transcription.language_model
: str
The language model used for transcription.acoustic_model
: str
The acoustic model used for transcription.language_code
: str
The language code representing the language of the transcription.status
: str
The status of the transcription.audio_url
: str
The URL of the audio file being transcribed.text
: str
The full transcription text.words
: Anyutterances
: Anyconfidence
: int | floataudio_duration
: Anypunctuate
: boolformat_text
: booldual_channel
: boolwebhook_url
: Anywebhook_status_code
: Anywebhook_auth
: boolwebhook_auth_header_name
: Anyspeed_boost
: boolauto_highlights_result
: Anyauto_highlights
: boolaudio_start_from
: Anyaudio_end_at
: Anyword_boost
: Anyboost_param
: Anyfilter_profanity
: boolredact_pii
: boolredact_pii_audio
: boolredact_pii_audio_quality
: Anyredact_pii_policies
: Anyredact_pii_sub
: Anyspeaker_labels
: boolcontent_safety
: booliab_categories
: boolcontent_safety_labels
: dictiab_categories_result
: dictlanguage_detection
: boolcustom_spelling
: Anythrottled
: Anyauto_chapters
: boolsummarization
: boolsummary_type
: Anysummary_model
: Anycustom_topics
: booltopics
: Anyspeech_threshold
: Anydisfluencies
: boolsentiment_analysis
: boolchapters
: Anysentiment_analysis_results
: Anyentity_detection
: boolentities
: Anysummary
: Anyspeakers_expected
: AnyUsing Text Anonymization, you can remove personal / private information from your tool.
Arguments
input
: str
The text you would like to processReturns
output
: strUsing NeMo Guard Rails content filter, you can ensure that your content is profanity free.
Arguments
input
: str
The text you would like to processusecase
: str
The usecase for which you want to moderate the output of you LLMdocuments
: Any
If you have selected Topic Control
, these are the texts that will be used.colang_content
: str
For examples, see the sample_rails
folder in any of these examples: https://github.com/NVIDIA/NeMo-Guardrails/tree/main/examplesyaml_content
: str
For examples, see the sample_rails
folder in any of these examples: https://github.com/NVIDIA/NeMo-Guardrails/tree/main/examplesReturns
output
: strInsert markdown notes
Arguments
markdown
: strReturns
Run a Relevance API request
Arguments
path
: str
Path to call for api methodmethod
: str
The HTTP method to use.body
: dict
The body to send with the request.Returns
response_body
: dictstatus
: int | floatbody
: strurl
: strresponse_headers
: dictFetch a linkedin profile with rapid API
Arguments
url
: str
The URL of the linkedin profile you’d like to scrapemethod
: str
The type of request to make to LinkedInReturns
data
: AnyPick from a variety of methods to extract a useful summary of information from knowledge.
Arguments
knowledge
: stroptimization
: dicttoken_limit
: int | float
The Summary generated will be limited to less than this many words.Returns
answer
: strcredits_cost
: int | floatoptimization
: dictUse a large multimodal model such as GPTV or Gemini-Vision
Arguments
prompt
: str
The prompt that is fed to the model.model
: str
The model to use for completion. If using gpt3.5, if you do not set your own API Key you will be charged credits for usage. For other models, Make sure to set an API key.images
: Any
The list of images that the model has acces to.temperature
: int | float
Temperature of the selected model. Typically, higher temperature means more random output.timeout_ms
: int | float
Timeout in milliseconds, request will timeout after this much time if it runs for too long.Returns
Convert a string to a JSON object.
Arguments
text
: str
The string to convert to JSON.Returns
output
: AnyExport data to a temporary downloadable file. It will be deleted after 1 hour automatically.
Arguments
data
: Any
Data to store in the file. This could be csv, json content etc.extension
: str
Extension of the file to save to. Defaults to csvReturns
export_url
: strFilter & extract data from JSON
Arguments
haystack
: dict
The JSON object to filter.needle
: Any
Fields to keep in the JSON object.Returns
filtered
: AnyFilter & extract data from string
Arguments
regex
: str
The regex to use for extraction, in JS flavor.string
: str
The string to search for matches.Returns
match
: Anyfound
: boolExecutes a transformation dynamically by a passed in ID. Useful for creating agents.
Arguments
transformation_id
: str
The ID of the transformation to run.params
: Any
The parameters to pass to the transformation.Returns
output
: AnyExecutes any one of your other tools.
Arguments
studio_id
: str
The tool to run.project
: str
The ID of the project the tool belongs to.region
: str
The ID of the region the tool belongs to.version
: strraise_subtool_errors
: bool
When true, errors from the subtool being called will also will also be raised in the current tool.params
: dict
The parameters to pass to the tool.studio_override
: dictwait_for_completion
: bool
If checked this step will wait until the tool completes, and return its output and errors. Otherwise this step will start the tool, and immediately return tool run information for polling.Returns
output
: dictstate
: dictstatus
: strerrors
: Anycost
: int | floatcredits_used
: AnyexecutionTime
: int | floatjob_info
: dictSplit text into an array of strings.
Arguments
text
: str
The text to split.method
: str
Separator will split the text by the separator. Tokens will split the text into chunks of a given number of tokens.num_tokens
: int | float
The number of tokens to split the text into.num_tokens_to_slide_window
: int | float
When using number of tokens, this param allow overlap between each sliced result, often improving accuracy for Ask Relevance.sep
: str
The separator to split the text by.Returns
chunks
: AnyCombine an array of strings into a single array.
Arguments
array
: Anymethod
: strReturns
combined
: AnyRun Python code
Arguments
code
: str
The python code to execute, returning the value to be used as the output of this transformation.backend
: str
Which backend service should run the code?packages
: Any
The python packages you would like to install to run your coderun_commands
: Any
A list of linux terminal commands to be run prior to executing codesession_id
: str
Define a custom session that saves all local variables between python executions. This only applies when the backend is Modal Labs.num_gpus
: str
Should the code use a GPU?
This only applies when the backend is Modal Labs.num_cpus
: str
How many CPU cores should the code use?
This only applies when the backend is Modal Labs.memory
: int | float
The Amount of Memory (RAM) in MBs the code will use.
This only applies when the backend is Modal Labs.timeout
: int | float
How long before the code times out in seconds?
This only applies when the backend is Modal Labs.raise_error
: str
Should python Step errorReturns
transformed
: Any
Return value of provided codeduration
: int | float
Duration of provided code in millisecondsstdout
: Any
Return stdout of provided codestderr
: Any
Return stderr of provided codesession_id
: str
The session ID used to run the executioncredits_cost
: int | float
Extra calculated credits based on image usedUpload to a temporary file in S3.
Arguments
blob
: Anyextension
: strReturns
url
: strRun Google search via serper.dev
Arguments
search_query
: str
Query to search forcountry_code
: str
Which region would you like to google search from?locale
: str
Which locale would you like to google search from?Returns
peopleAlsoAsk
: AnyrelatedSearches
: Anyorganic
: Anyuser_key_used
: boolTag or extract information from images
Arguments
image_url
: str
URL of the image to analyse. JPEG/JPG, PNG and WebP are supported.prompt
: str
The question(s) to answer about the image.Returns
output
: strGenerate image via Stable Diffusion
Arguments
prompt
: str
What do you want to see? You can use a single word or a full sentence.num_outputs
: int | float
Select the number of images you would like to generate.Returns
images
: AnyGenerate Speech from text
Arguments
prompt
: str
What do you want to hear? You can use a single word or a full sentence.voice
: strReturns
audio_url
: strGenerate audio from text with ElevenLabs
Arguments
prompt
: str
What do you want to hear? You can use a single word or a full sentence.voice_id
: strmodel_id
: strpronounciation_dictionary_locators
: Anysimilarity_boost
: int | float
Voice settings: similarity booststability
: int | float
Voice settings: stabilitystyle
: int | float
Voice settings: styleuse_speaker_boost
: bool
Voice settings: speaker boostReturns
audio_url
: strSend an AI phone call with Bland.AI
Arguments
phone_number
: str
The phone number to call. Country code defaults to +1 (US) if not specified. Formatting is flexible, however for the most predictable results use the E.164 format.task
: str
Provide instructions, relevant information, and examples of the ideal conversation flow.transfer_list
: dict
Give your agent the ability to transfer calls to a set of phone numbers. Overrides transfer_phone_number if a transfer_list.default is specified. Will default to transfer_list.default, or the chosen phone numbermodel
: str
Select a model to use for your call. In nearly all cases, enhanced is the best choice for now.transfer_phone_number
: str
A phone number that the agent can transfer to under specific conditions - such as being asked to speak to a human or supervisor.answered_by_enabled
: bool
Enables machine detection when the call starts to determine whether the call was answered by a person or a voicemail.from
: str
Specify a purchased Outbound Number to call from. Country code is required, spaces or parentheses must be excluded. By default, calls are initiated from a separate pool of numbers owned by Bland.voice_id
: str
Determines the voice of the AI agent, in conjunction with reduce_latency.voice_preset_id
: str
Use a voice preset instead of specifying individual voice settings.start_time
: str
The time you want the call to start. If you don’t specify a time (or the time is in the past), the call will send immediately. Set your time in the format YYYY-MM-DD HH:MM:SS -HH:MM (ex. 2021-01-01 12:00:00 -05:00). The timezone is optional, and defaults to UTC if not specified.webhook
: str
The webhook should be a http / https callback url. We will send the call_id and transcript to this URL after the call completes. This can be useful if you want to have real time notifications when calls finish.wait_for_greeting
: bool
Should the AI speak first or wait for someone else to talk? Creates more realistic conversations when answered with “Hello?”, “This is {name} speaking.” and so on.first_sentence
: str
A phrase that your call will start with instead of a generating one on the fly. This works both with and without wait_for_greeting. Can be more than one sentence, but must be less than 200 characters.record
: bool
To record your phone call, set record to true. When your call completes, you can access the recording by requesting the /call/recording endpoint.language
: str
Select a supported language of your choice. Optimizes every part of our API for that language - transcription, speech, and other inner workings.max_duration
: int | float
Set the longest you want the call to possibly go in minutes. After the max_duration minutes have passed, the call will automatically end.amd
: bool
AMD mode helps our AI navigate phone trees and IVR systems. If you know your call will hit an automated system you should switch it on. NOTE: AMD mode causes increased delay for the first response, even if answered by a human. Highly recommended to set to false in the majority of cases.interruption_threshold
: dictReturns
status
: strcall_id
: strRetrieve the call details for a Bland AI call
Arguments
call_id
: str
The call id to fetch the transcript for.Returns
response
: dictGenerate professional avatar videos with HeyGen
Arguments
prompt
: str
What do you want to hear? You can use a single word or a full sentence.voice_id
: stravatar_id
: stravatar_style
: strscale
: int | floatoffset_x
: int | floatoffset_y
: int | floatbackground
: strratio
: strtest
: bool
Set the test flag to true to use test mode. Generation in test mode will not cost your credits and will contain a watermark on your video.Returns
video_url
: strGenerate professional avatar videos with D-ID
Arguments
prompt
: str
What do you want to hear? You can use a single word or a full sentence.voice_id
: stravatar_id
: strReturns
video_url
: strGenerate image via Stable Diffusion XL
Arguments
prompt
: Any
What do you want to see? You can use a single word or multiple sentences.negative_prompt
: Any
What do you not want to see? You can use a single word or multiple sentences.height
: int | float
What should the height of the generated images be?width
: int | float
What should the height of the generated images be?guidance_scale
: int | float
Higher guidance scale encourages to generate images that are closely linked to the text prompts usually at the expense of lower image qualitynum_images_per_prompt
: int | float
How many images do you want to generate per prompt?eta
: int | float
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502.steps
: int | float
The number of denoising steps. More denoising steps usually lead to a higher quality image at the expense of slower inference.Returns
images
: Anycredits_used
: int | floatTruncate text based on a specified token count.
Arguments
text
: Any
The text to truncate.num_tokens
: int | float
The number of tokens to truncate the text or array by.Returns
chunks
: AnyScrape and access website content
Arguments
website_url
: str
URL of the website to scrape.method
: strmodel
: dict
Type of service to use when scraping the websiteReturns
output
: dictoutputs
: Anyuser_key_used
: boolcredits_cost
: int | floatSend a Slack message
Arguments
oauth_account_id
: str
The Slack channel to post message to.message
: strblocks
: AnyReturns
status
: int | floatSend a WhatsApp message
Arguments
from_phone_number_id
: str
The from phone number ID (not to be confused with the actual phone number)to
: str
The phone number to messagemessage
: strReturns
status
: int | floatcontact_id
: strdetails
: strSend a WhatsApp template message
Arguments
from_phone_number_id
: str
The from phone number ID (not to be confused with the actual phone number)to
: str
The phone number to messagetemplate_name
: str
The name of the template you wish to sendtemplate_language
: str
The language for this templatecontent_parameters
: dictReturns
status
: int | floatcontact_id
: strdetails
: strCreate a Zendesk Ticket
Arguments
email
: str
The email address you use to access your Zendesk accountsubdomain
: str
https://{subdomain}.zendesk.com/ticket
: strReturns
status
: int | floatCreate a linear ticket
Arguments
oauth_account_id
: str
The Linear Account used to create the ticket.team
: strtitle
: strdescription
: strproject
: strpriority
: str
The priority of the linear ticketteamLabels
: Any
The labels specific to the Linear team.workspaceLabels
: Any
The labels specific to the Linear workspace.Returns
status
: int | floatMake an authenticated request to an Apollo API using credentials saved in Relevance
Arguments
method
: str
The HTTP method to use.path
: str
Path to make the request to. e.g. /v1/auth/healthbody
: dict
The body to send with the request.Returns
response_body
: Anystatus
: int | floatbody
: strurl
: strresponse_headers
: dictSearch for users on apollo for people enrichment. Makes an authenticated request to Apollo API using credentials saved in Relevance
Arguments
searchType
: dict
How you want to retrieve your information from Apollo.Returns
response_body
: Anystatus
: int | floatbody
: strurl
: strresponse_headers
: dictExecute postgres SQL. Setup details in the api keys section.
Arguments
sql
: str
The SQL Query you want to performrejectUnauthorized
: bool
Whether to verify the server’s identity. Strongly suggested to be true outside of developmentReturns
rows
: Anymetadata
: dictMake an authenticated request to a Hubspot API using credentials saved in Relevance
Arguments
method
: str
The HTTP method to use.oauth_account_id
: str
The Hubspot account.path
: str
Path to make the request to. e.g. /v1/auth/healthbody
: dict
The body to send with the request.Returns
response_body
: Anystatus
: int | floatbody
: strurl
: strresponse_headers
: dictMake an authenticated request to a Freshdesk REST API using your api key
Arguments
freshdesk_portal_domain
: str
The domain of your freshdesk portal. e.g. relevance.freshdesk.commethod
: str
The HTTP method to use.path
: str
Path to make the request to. e.g. /api/v2/ticketsbody
: dict
The body to send with the request.Returns
response_body
: Anystatus
: int | floatbody
: strurl
: strresponse_headers
: dictTranslate 100+ languages via Google Translate
Arguments
text
: str
Text to translate.sourceLanguageCode
: str
Language of the input text. Determined automatically if no language is selected.targetLanguageCode
: str
Language to translate the input text to.Returns
result
: strExtract text from an Excel document.
Arguments
excel_url
: str
The URL of the Excel document to convert to text.Returns
text
: AnyExtract text from a Word document.
Arguments
doc_url
: str
The URL of the Word document to convert to text.Returns
text
: strExtract text from a file. Supports JSON, CSV, PDF and Word files.
Arguments
llm_friendly_format
: bool
Set to false to not transform text into an llm friendly format.file_url
: str
The URL of the file to convert to text.Returns
text
: strdetected_type
: strmetadata
: dictThis transformation is useful for testing and debugging. It simply returns whatever you input.
Arguments
data
: AnyReturns
data
: AnyConverts an XLSX or CSV file to a list of JSON objects, with one object per row.
Arguments
file_url
: strReturns
rows
: AnySearch for a string in a document and return a match with start and end indices if found.
Arguments
query
: str
The substring to search for in the text. Uses fuzzy search.text
: str
The text to search in.Returns
text
: strmatch
: strstart
: int | floatend
: int | floatMake an authenticated request to a Google API using credentials saved in Relevance
Arguments
oauth_account_id
: str
The Google account to use for authentication. Any API requests will be made on behalf of this account.oauth_permission_type
: strurl
: str
Full url to make the request to. e.g. https://gmail.googleapis.com/gmail/v1/users/me/messagesmethod
: str
The HTTP method to use.body
: dict
The body to send with the request.Returns
response_body
: dictstatus
: int | floatbody
: strurl
: strresponse_headers
: dictMake an authenticated request to a Salesforce REST API using credentials saved in Relevance
Arguments
oauth_account_id
: str
The Salesforce account to use for authentication. Any API requests will be made on behalf of this account.api_version
: str
The Salesforce REST API version to use e.g. 59.0. By default we will use the latest version.path
: str
Path to make the request to. e.g. /sobjects/Accountmethod
: str
The HTTP method to use.body
: dict
The body to send with the request.Returns
response_body
: dictstatus
: int | floatbody
: strurl
: strresponse_headers
: dictMake an authenticated request to a Microsoft Graph API endpoint using credentials saved in Relevance
Arguments
oauth_account_id
: str
The Microsoft account to use for authentication. Any API requests will be made on behalf of this account.oauth_permission_type
: strpath
: str
Path to make the request to. e.g. /v1.0/me/messages. See [Microsoft’s Graph Explorer](https://developer.microsoft.com/en-us/graph/graph-explorer/) for a list of endpointsmethod
: str
The HTTP method to use.body
: dict
The body to send with the request.headers
: dict
The headers to send with the request. The ‘Authorization’ header is always sent and cannot be overwritten.Returns
response_body
: Anystatus
: int | floatbody
: strurl
: strresponse_headers
: dictSend an email using a Google / Gmail account stored in Relevance
Arguments
oauth_account_id
: str
The Gmail account from which to send emails.to
: Any
The email address to send the email to.cc
: Any
The email addresses to cc.bcc
: Any
The email addresses to blind cc.subject
: str
The subject of the email.body
: str
The text or html body of the email.gmail_thread_id
: str
The gmail thread in which to send this message.reply_to_message
: str
The message id to reply to. If blank, the email will be sent as a new message.draft
: bool
If true, the email will be saved as a draft instead of sent immediately.Returns
gmail_thread_id
: strgmail_message_id
: strmessage_id
: strraw_email
: strSend a message to an agent
Arguments
agent_id
: str
The agent to messageagent_override
: dictoverride
: dictmessage
: str
The text message to send to the agent.conversation_id
: str
The conversation to send the message to. If blank, a new conversation will be created.wait_for_reply
: bool
Deprecated - If true, this step will wait until the agent replies before continuing. Otherwise, the message will be sent, and the step will continue immediately.metadata
: dict
Information to store in message metadata.params
: dict
Params to set in the agent before runningReturns
job_info
: dictconversation_id
: stragent_id
: strstate
: strmessage
: dictCheck Google calendar availability
Arguments
oauth_account_id
: str
The Google account of the calendar to check.start_date
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00end_date
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00slot_duration
: int | float
Duration of booking in minutesstart_working_time
: strend_working_time
: strtimezone
: strcalendars
: Any
If not defined, the Google account’s primary calendar is usedReturns
calendars
: dictCreate a calendar event using a Google account stored in Relevance
Arguments
oauth_account_id
: str
The Google account of the calendar to check.event_title
: strevent_description
: strstart_datetime
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00end_datetime
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00attendees
: Anycalendar
: str
If not defined, the Google account’s primary calendar is usedlocation
: str
Location for the meeting. Can be a physical location, or a link to a virtual meetingguests_can_modify
: bool
Boolean that gives event guests the ability to modify the eventReturns
event_id
: strSend an email using a Microsoft account stored in Relevance
Arguments
oauth_account_id
: str
The Microsoft account from which to send emails.to
: Any
The email address to send the email to.cc
: Any
The email addresses to cc.bcc
: Any
The email addresses to blind cc.subject
: str
The subject of the email.body
: str
The text or html body of the email.reply_to_message
: str
The outlook message id to reply to. If blank, the email will be sent as a new message.draft
: bool
If true, the email will be saved as a draft instead of sent immediately.Returns
outlook_thread_id
: stroutlook_message_id
: strmessage_id
: strraw_email
: strCheck calendar availability using a Microsoft Outlook account stored in Relevance
Arguments
oauth_account_id
: str
The Microsoft account from which to send emails.start_date
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00end_date
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00slot_duration
: int | float
Duration of booking in minutesstart_working_time
: strend_working_time
: strtimezone
: strcalendars
: Any
Calendars must be a Microsoft 365 work or school email in your organisationis_organizer_optional
: bool
If there are no calendars to query, organizer is required by defaultReturns
calendars
: dictCreate a calendar event using a Microsoft Outlook account stored in Relevance
Arguments
oauth_account_id
: str
The Microsoft account from which to send emails.subject
: strbody
: strstart_datetime
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00end_datetime
: str
ISO formatted string for start date. e.g. 2023-12-06T19:00:00+00:00attendees
: Anycalendar
: str
If not defined, the Microsoft account’s primary calendar is usedlocation
: str
Location for the meeting. Can be a physical location, or a link to a virtual meetingReturns
event_id
: strDelay for specified amount of time.
Arguments
delay
: int | float
Delay amount in secondsReturns
delay
: int | floatRetrive data from your dataset.
Arguments
knowledge_set
: strinclude_fields
: Anypage_size
: int | floatReturns
results
: AnyMake a call to the Zoom API
Arguments
oauth_account_id
: strpath
: str
Path to make the request to, and must include the API version. e.g. /v2/users/memethod
: str
The HTTP method to use.body
: dict
The body to send with the request.Returns
response_body
: dictstatus
: int | floatbody
: strurl
: strresponse_headers
: dictMake a phone call with AI
Arguments
phone_number
: strassistant_model
: strassistant_system_prompt
: strassistant_prompt
: strassistant_first_message
: strassistant_final_message
: strvoice_id
: str
What kind of spoken voice do you want your assistant to have?configuration
: dict
See https://docs.vapi.ai/api-reference/calls/create-phone-call for total list of parameterswait_for_finish
: bool
Do we want this step to return once the call has been sent or once the call has concluded?Returns
status
: strcall_id
: strcredits_used
: int | floatMake a phone call with AI using Vapi AI
Arguments
phone_number_id
: strassistant_id
: strcustomer_phone_number
: str
See https://docs.vapi.ai/api-reference/calls/create-phone-call for total list of parametersphone_number
: dict
See https://docs.vapi.ai/api-reference/calls/create-phone-call for total list of parametersassistant
: dict
See https://docs.vapi.ai/api-reference/calls/create-phone-call for total list of parameterscustomer
: dict
See https://docs.vapi.ai/api-reference/calls/create-phone-call for total list of parameterswait_for_finish
: bool
Do we want this step to return once the call has been sent or once the call has concluded?Returns
status
: strcall_id
: strcredits_used
: int | floatRetrieve the call details for a call made with Vapi AI
Arguments
use_user_key
: boolcall_id
: str
The call id to fetch the transcript for.Returns
response
: dictcodes Art & Legible Text Art directly in your products. Request a free API key to get started.
Arguments
qrContent
: strprompt
: strworkflow
: strqrStrength
: int | floatqrImage
: strqrSteps
: int | floatqrCodePattern
: str
https://docs.quickqr.art/user-guide/qr-pattern-listseed
: int | floatnegativePrompt
: strReturns
image_url
: strresponse_body
: dictSearch linkedin profiles with rapid API
Arguments
geo_codes
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencesgeo_codes_exclude
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencestitle_keywords
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencestitle_keywords_exclude
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencescurrent_company_ids
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencespast_company_ids
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencesfunctions
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referenceskeywords
: str
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referenceslimit
: int | float
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencesReturns
data
: AnySearch linkedin companies with rapid API
Arguments
company_headcounts
: strheadquarters_location
: Anyindustry_codes
: Anyhiring_on_linkedin
: boolrecent_activities
: Any
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referenceskeywords
: str
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referenceslimit
: int | float
https://rapidapi.com/freshdata-freshdata-default/api/fresh-linkedin-profile-data/tutorials/search-employees-endpoint-payload-referencesReturns
data
: AnySearch LinkedIn Jobs with rapid API
Arguments
keywords
: str
Keywords to search for in job postings.geo_code
: str
Param to target jobs in specific region/country. To search worldwide, use 92000000. Find other geo codes via the provided link.date_posted
: str
Filter by the date when jobs were posted.experience_level
: str
Filter by the level of experience required for the job.company_id
: int | float
Filter by the ID of the company posting the job.title_id
: str
Filter by the ID of the job title. Find title IDs via the provided link.onsite_remote
: str
Filter by the working arrangement of the job.function_id
: str
Get the function ID of your choice following the provided instruction.industry_code
: str
Filter by the industry code of the job.sort_by
: str
Specify the sorting criteria for the job listings.start
: str
Specify the starting point for retrieving job listings.easy_apply
: str
Filter by whether the job offers easy application process.under_10_applicants
: bool
Filter by whether the job has less than 10 applicants.Returns
data
: AnyChoose which of your collections to list.
Arguments
collection_id
: strcms_locale_id
: stroffset
: int | floatlimit
: int | floatReturns
Create or Update a webflow collection item
Arguments
collection_id
: stritem_id
: strcms_locale_id
: stris_archived
: boolis_draft
: boolfield_data
: dictReturns
Get a webflow collection item
Arguments
collection_id
: stritem_id
: strcms_locale_id
: strReturns
Delete a webflow collection item
Arguments
collection_id
: strcms_locale_id
: stritem_id
: strReturns
Make an authenticated request to a Salesforce REST API using credentials saved in Relevance
Arguments
oauth_account_id
: str
The Salesforce account to use for authentication. Any API requests will be made on behalf of this account.ids_to_convert
: AnyReturns