Save to memory
Module · Stable
Store an uploaded file, inline base64 bytes, OR raw text durably as a memory item, optionally linked to a source document.
Persists a file (file_path), inline bytes (content_base64 + filename — e.g. a photo an edge device delivered inside a webhook JSON payload), or raw text (text — written to a .txt file internally, e.g. a transcript or an AI-generated summary) to durable storage and returns a stable document_id plus the stored_path. Exactly one source is required; when several are set, file_path wins, then content_base64, then text. Pass parent_doc_id (another document_id from an earlier store step) to link a derived artefact — e.g. a transcript — back to its source recording. Use before OCR or Compress steps when the source file is an actor-uploaded attachment that must survive run retries.
When to use
Section titled “When to use”Use when an Ask Actor step, a trigger, or a Read file (storage) step has produced a raw file_path and you need a stable document_id before passing the file to OCR or Compress — or when a prior step (e.g. Speech to Text, Ask AI) produced text you want saved as its own memory item. The document_id it emits is the durable handle every downstream processing step expects. Set parent_doc_id to the source document_id when storing something derived from an already-stored file, so the two stay linked.
When not to use
Section titled “When not to use”If you only need to read text from a document that is already stored, go directly to OCR — it accepts a file_path without a prior store step. To reduce file size before storing, place Compress ahead of this step rather than replacing it.
Inputs
Section titled “Inputs”Configured per use: file_path, content_base64, filename, text, document_date, document_type, actor_id, parent_doc_id.
Outputs
Section titled “Outputs”Auto-generated from the skill registry (load_skills()). Do not edit by hand.