create batch
This block is part of the anthropic group and was last modified in core v15.0.0.
create batch v15.0.0 Submit many prompts as one asynchronous batch at 50% of the standard price (results within 24h). Wire on_finished/on_item to be called automatically when the batch completes (the system polls in the background), or poll the batch_id yourself with the get batch block.
- flow
flow - connection required
anthropicconnection - requests required
array<object>One object per prompt: {prompt (required), system?, model?, max_tokens?, custom_id? (your key to match results)}. custom_id defaults to request-N.
- poll interval (s)
numberSeconds between background poll sweeps when on_finished/on_item are wired. Default 120, minimum 30. • Hidden when unused
- flow
flow - connection
anthropicconnection - success
condition - batch id
valuePoll this with the get batch block.
- status
valuein_progress | canceling | ended.
- counts
objectprocessing, succeeded, errored, canceled, expired.
- count
numberNumber of requests submitted.
- errors
array<value>Error codes when the call fails (success=false): config_error, internal_error, model_not_allowed, transport_error.
- error detail
value - on finished
execute_functionRuns in the background when the batch completes (the system polls for you). Receives the per-result summary; results carries one entry per request.
- on item
execute_functionRuns in the background once per finished request (custom_id, status, text, usage, cost). Optional; use it to process results one by one.
batch |
bulk message |
async batch |
mass prompt |
50% off |
A typical wiring for create batch: a function trigger starts the flow; with values feeds the requests pin; the function out pin wires to an execute block.
Version history
Introduced in v15.0.0.
Back to
anthropic Return to the main group to view all sub-groups Back to anthropic Return to the group to view all blocks within this group