Source files & AI context
How to upload files, what formats are supported, and how the AI processes them.
Source files belong to a WritersRoom project. Upload the current draft and supporting documents, then check the extraction status and available text before asking the AI to use them.
WritersRoom source formats
| Format | What is read |
|---|---|
| Extractable text. Scanned pages and embedded images may not provide usable text. | |
| DOCX | Extracted document text; visual layout and embedded images are not a full document render. |
| TXT | Plain text. |
| MD | Text including Markdown syntax. |
Uploading a file, extracting its text and selecting text for an AI request are different steps. File size is measured in bytes; context ceilings below are measured in characters, not words or pages. Each workflow selects a bounded amount of usable text. The assistant, Coverage, Development Intelligence, Consistency Check and memory extraction can read different amounts. A stored document is not a promise that every character is included in every request.
Use files effectively
Check extraction
If a file has little or no readable text, use a text-based export. A successful upload alone does not establish a successful analysis.
Choose sources for the task
Use the correct screenplay and any references offered by that workflow. Context selection and reference opt-in are not the same as uploading another copy.
Read the result in context
A report or answer can only use the selected, readable material within its own limits. For long drafts, ask about a specific section or use the appropriate report workflow.
Uploads and AI actions
An upload that starts automatic AI analysis uses one action under the applicable AI-access rules. An authorised standalone report purchase and its required source analysis do not also debit a subscription's 500-action allowance. Manual writing and reopening saved output are separate from new AI generation.