Skip to main content

CLI Reference

The DocuDevs CLI provides convenient command-line access to all platform functionality. The CLI is included with the SDK installation and supports both high-level convenience commands and low-level API operations.

Installation​

The CLI is included when you install the DocuDevs SDK:

pip install docu-devs-api-client

Authentication​

The CLI supports multiple authentication methods:

# Primary environment variable
export DOCUDEVS_TOKEN=your-api-key-here

# Legacy support (also works)
export API_KEY=your-api-key-here

# Use commands without --token
docudevs process document.pdf

Command Line Option​

# Pass token with each command
docudevs --token=your-api-key-here process document.pdf

High-Level Commands​

These commands handle the complete workflow for common operations.

process​

Upload and process a document with AI extraction in one command.

docudevs process document.pdf [OPTIONS]

Options:

  • --prompt TEXT: Extraction prompt describing what to extract
  • --prompt-file PATH: Read extraction prompt from a text file
  • --schema TEXT: JSON schema for structured extraction
  • --schema-file PATH: Load JSON schema from file
  • --mime-type TEXT: Override MIME type detection for the input file
  • --configuration TEXT: Process using a named configuration
  • --ocr [DEFAULT|NONE|PREMIUM|AUTO|EXCEL]: OCR processing type (default: DEFAULT)
  • --llm [DEFAULT|MINI|HIGH]: LLM model to use (default: DEFAULT)
  • --barcodes: Enable barcode and QR detection for this run
  • --extraction-mode [OCR|SIMPLE|STEPS|WORKBOOK]: Force a specific extraction pipeline. WORKBOOK requires a nonblank --prompt, rejects --schema, and only accepts an .xlsx input file.
  • --describe-figures: Request figure descriptions when supported
  • --timeout INTEGER: Timeout in seconds (default: 60)
  • --wait/--no-wait: Wait for processing to complete (default: wait)
  • --tool JSON: Attach a tool descriptor as JSON or @path (must include type; can be used multiple times)

Examples:

# Extract invoice data
docudevs process invoice.pdf --prompt="Extract invoice data including total, date, vendor"

# Use JSON schema for structured extraction
docudevs process contract.pdf --schema='{"type": "object", "properties": {"party1": {"type": "string"}}}'

# Use premium OCR and don't wait for result
docudevs process scan.pdf --ocr=PREMIUM --no-wait

# Read prompt and schema from files
docudevs process invoice.pdf --prompt-file instructions.txt --schema-file schema.json

# Process using a saved configuration and enable barcode detection
docudevs process receipt.pdf --configuration retail-config --barcodes

process-map-reduce​

Upload and process a document using map-reduce chunking.

docudevs process-map-reduce document.pdf [OPTIONS]

Options:

  • --prompt/--prompt-file: Extraction instructions (same as process)
  • --schema/--schema-file: JSON schema (same as process)
  • --mime-type TEXT: Override MIME type detection
  • --ocr [DEFAULT|NONE|PREMIUM|AUTO|EXCEL]: OCR processing type (default: DEFAULT)
  • --llm [DEFAULT|MINI|HIGH]: LLM model to use (default: DEFAULT)
  • --barcodes: Enable barcode/QR detection
  • --extraction-mode [OCR|SIMPLE|STEPS|WORKBOOK]: Pipeline override. WORKBOOK is rejected for map-reduce processing.
  • --describe-figures: Request figure descriptions when supported
  • --tool JSON: Attach a tool descriptor as JSON or @path (must include type; can be used multiple times)
  • --pages-per-chunk INTEGER: Pages per chunk (default: 1)
  • --overlap INTEGER: Overlap between chunks (default: 0)
  • --dedup-key TEXT: JSON path used to deduplicate rows across overlapping chunks
  • --split-type [page|markdown-header]: Chunking mode (default: page)
  • --split-header-level INTEGER: Header level (1 or 2) when --split-type markdown-header
  • --header-page-limit INTEGER: Number of pages reserved for header extraction
  • --header-include-in-rows: Include header pages in row processing
  • --header-row-prompt-augmentation TEXT: Extra context injected into each chunk prompt
  • --header-schema/--header-schema-file: Header-specific schema
  • --header-prompt TEXT: Header-specific prompt
  • --header-page-index INTEGER: Repeatable, explicit header page indices
  • --stop-when-empty: Stop processing after empty chunks (use with --empty-chunk-grace)
  • --empty-chunk-grace INTEGER: Number of empty chunks tolerated when --stop-when-empty is set
  • --timeout INTEGER: Wait timeout in seconds (default: 60)
  • --wait/--no-wait: Wait for processing to complete (default: wait)

When using --split-type markdown-header, --overlap and --dedup-key are not supported.

Example:

docudevs process-map-reduce large-invoice.pdf \
  --prompt "Extract line items" \
  --schema-file schema.json \
  --pages-per-chunk 3 \
  --overlap 1 \
  --dedup-key "lineItems.sku" \
  --header-page-limit 1 \
  --header-schema '{"invoiceNumber":"string"}'

# Split by markdown subsection (##)
docudevs process-map-reduce handbook.pdf \
  --prompt "Extract policies by section" \
  --split-type markdown-header \
  --split-header-level 2

ocr-only​

Upload and process a document with OCR-only mode (no AI extraction).

docudevs ocr-only document.pdf [OPTIONS]

Options:

  • --ocr [DEFAULT|NONE|PREMIUM|AUTO|EXCEL]: OCR processing type (default: DEFAULT)
  • --format [plain|markdown|jsonl]: Output format (default: plain, or jsonl for EXCEL)
  • --timeout INTEGER: Timeout in seconds (default: 60)
  • --wait/--no-wait: Wait for processing to complete (default: wait)

Examples:

# Basic OCR
docudevs ocr-only document.pdf

# OCR with markdown formatting
docudevs ocr-only document.pdf --format=markdown

# Premium OCR
docudevs ocr-only document.pdf --ocr=PREMIUM

wait​

Wait for a job to complete and return the result.

docudevs wait JOB_GUID [OPTIONS]

Options:

  • --timeout INTEGER: Timeout in seconds (default: 60)

Example:

docudevs wait 550e8400-e29b-41d4-a716-446655440000

Batch Processing​

Upload and process multiple documents as one batch job.

batch process​

docudevs batch process FILE [FILE ...] [OPTIONS]

Options:

  • --prompt/--prompt-file: Extraction instructions (same as process)
  • --schema/--schema-file: JSON schema (same as process)
  • --mime-type TEXT: Explicit MIME type for all documents
  • --ocr [DEFAULT|NONE|PREMIUM|AUTO|EXCEL]: OCR processing type
  • --llm [DEFAULT|MINI|HIGH]: LLM model to use
  • --barcodes: Enable barcode and QR code detection
  • --extraction-mode [OCR|SIMPLE|STEPS|WORKBOOK]: Extraction mode override. WORKBOOK is rejected for batch processing.
  • --describe-figures: Request figure descriptions when supported
  • --max-concurrency INTEGER: Maximum concurrent document processing within the batch
  • --format [json|csv|excel]: Output format for results (implies --wait)
  • --output PATH: Save results to this file (auto-generated from --format if omitted)
  • --configuration TEXT: Named configuration for configuration-backed batch processing. Cannot be combined with --prompt, --schema, --ocr, --llm, --barcodes, --extraction-mode, or --describe-figures.
  • --lookup-file PATH: Batch-scoped lookup file that overrides any configuration lookup file for this batch
  • --timeout INTEGER: Timeout in seconds (default: 300)
  • --wait/--no-wait: Wait for processing to complete (default: wait)

Examples:

# Batch-process several invoices with a shared prompt
docudevs batch process invoice-1.pdf invoice-2.pdf invoice-3.pdf \
  --prompt "Extract invoice data" \
  --max-concurrency 5

# Batch-process using a saved configuration and export to Excel
docudevs batch process *.pdf --configuration invoice-config --format excel --output batch-results.xlsx

Configuration Management​

Manage named processing configurations for reusable workflows.

list-configurations​

List all saved configurations.

docudevs list-configurations

get-configuration​

Get details of a specific configuration.

docudevs get-configuration CONFIG_NAME

save-configuration​

Save a processing configuration from a JSON file.

docudevs save-configuration CONFIG_NAME config.json

Example config.json:

{
  "prompt": "Extract invoice data",
  "schema": "{\"type\": \"object\"}",
  "ocr": "DEFAULT",
  "llm": "DEFAULT"
}

delete-configuration​

Delete a saved configuration.

docudevs delete-configuration CONFIG_NAME

configuration lookup-file upload​

Attach a lookup file to a named configuration. It applies to every job that uses the configuration until replaced or deleted.

docudevs configuration lookup-file upload CONFIG_NAME vendors.csv

configuration lookup-file delete​

Remove the lookup file attached to a configuration.

docudevs configuration lookup-file delete CONFIG_NAME

Template Management​

Work with document templates for form filling and extraction.

list-templates​

List all available templates.

docudevs list-templates

upload-template​

Upload a template document.

docudevs upload-template TEMPLATE_NAME template.pdf

template-metadata​

Fetch metadata for a template.

docudevs template-metadata TEMPLATE_NAME

For newly uploaded PDF templates, metadata extraction is asynchronous. If the template was just uploaded, wait briefly and run the command again.

delete-template​

Delete a template.

docudevs delete-template TEMPLATE_NAME

fill​

Fill a template with data from a JSON file and optionally write the result to disk.

docudevs fill TEMPLATE_NAME data.json --output filled.pdf

Example data.json:

{
  "fields": {
    "name": "John Doe",
    "address": "123 Main St",
    "date": "2024-01-15"
  }
}

Case Management​

Manage cases and documents stored within them.

cases list​

List all cases for the current organization.

docudevs cases list

cases create​

Create a new case.

Options:

  • --name TEXT: Case name (required)
  • --description TEXT: Optional case description
docudevs cases create --name "Quarterly Invoices" --description "Q4 2024 invoice processing"

cases get​

Retrieve details for a specific case.

docudevs cases get CASE_ID

cases update​

Update an existing case.

docudevs cases update CASE_ID --name "Quarterly Invoices (Updated)"

cases delete​

Delete a case and all associated documents.

docudevs cases delete CASE_ID

cases upload-document​

Upload a document into a case.

Options:

  • --filename TEXT: Custom filename for the uploaded document (defaults to the source file's name)
  • --mime-type TEXT: Explicit MIME type for the document (defaults to detection from the filename)
  • --metadata TEXT: Document metadata as a JSON object or comma-separated key=value pairs
docudevs cases upload-document CASE_ID invoice.pdf --metadata '{"department": "finance"}'

cases list-documents​

List documents stored in a case.

docudevs cases list-documents CASE_ID --page 0 --size 20

cases get-document​

Get metadata for a document stored in a case.

docudevs cases get-document CASE_ID DOCUMENT_ID

cases delete-document​

Remove a document from a case.

docudevs cases delete-document CASE_ID DOCUMENT_ID

cases get-summary​

Get the AI-generated summary for a document in a case.

docudevs cases get-summary CASE_ID DOCUMENT_ID

cases list-summaries​

List AI-generated summaries for all documents in a case.

docudevs cases list-summaries CASE_ID

Knowledge Base Management​

Manage knowledge bases derived from cases.

knowledge-base list​

List all knowledge bases.

docudevs knowledge-base list

knowledge-base add​

Promote a case to a knowledge base.

docudevs knowledge-base add CASE_ID

knowledge-base get​

Get a knowledge base by case ID.

docudevs knowledge-base get CASE_ID

knowledge-base remove​

Demote a case from knowledge base status.

docudevs knowledge-base remove CASE_ID

Contract Analysis​

Manage a knowledge base's contract-analysis profile, generations, test suite, trials, and immutable analysis runs. All commands are nested under knowledge-base contract-analysis. See Contract Analysis for the profile/draft/publish lifecycle and the run/review model.

knowledge-base contract-analysis get​

Get a contract-analysis profile draft or its published version.

docudevs knowledge-base contract-analysis get CASE_ID --view draft
  • --view [draft|published]: Which version to fetch (default: draft)

knowledge-base contract-analysis generate​

Start profile compilation from the eligible knowledge-base documents.

docudevs knowledge-base contract-analysis generate CASE_ID

knowledge-base contract-analysis generation-compare​

Show the previous, current, and proposed values for a regeneration.

docudevs knowledge-base contract-analysis generation-compare CASE_ID GENERATION_JOB_GUID

knowledge-base contract-analysis generation-apply​

Apply explicit per-element decisions against the latest comparison fingerprint.

docudevs knowledge-base contract-analysis generation-apply CASE_ID GENERATION_JOB_GUID --decisions decisions.json
  • --decisions PATH: JSON object mapping each changed lineage ID to an explicit action and optional reason (required)

knowledge-base contract-analysis update​

Replace the draft profile at the observed revision.

docudevs knowledge-base contract-analysis update CASE_ID --draft-revision 3 --profile profile.json
  • --draft-revision INTEGER: Observed draft revision (required)
  • --profile PATH (aliases --profile-file, --input): Draft profile JSON file (required)

knowledge-base contract-analysis publish​

Publish a validated draft profile at the observed revision.

docudevs knowledge-base contract-analysis publish CASE_ID --draft-revision 3 --acknowledge-warning missing_termination_clause
  • --draft-revision INTEGER: Observed draft revision (required)
  • --acknowledge-warning (aliases --acknowledge-warning-code, --warning-code): Warning code to acknowledge (repeatable)
  • --trial-attestation-id TEXT: Optional trial attestation ID backing the publish

knowledge-base contract-analysis suite-get​

Read the current immutable contract-analysis test suite.

docudevs knowledge-base contract-analysis suite-get CASE_ID

knowledge-base contract-analysis suite-update​

Create the next suite revision from a JSON array of test cases.

docudevs knowledge-base contract-analysis suite-update CASE_ID --revision 2 --cases cases.json
  • --revision INTEGER: Observed suite revision (required)
  • --cases PATH: JSON array of test cases (required)

knowledge-base contract-analysis trial​

Launch an explicit draft-profile trial and print its trial/run identities.

docudevs knowledge-base contract-analysis trial CASE_ID \
  --anchor-job-guid JOB_GUID \
  --idempotency-key trial-1 \
  --kind scored
  • --anchor-job-guid TEXT: Anchor job GUID (required)
  • --idempotency-key TEXT: Idempotency key (required)
  • --suite-revision INTEGER: Optional suite revision to trial against
  • --kind [exploratory|scored]: Trial kind (default: exploratory)
  • --suite-case-id TEXT: Optional single suite case to trial

knowledge-base contract-analysis trial-get​

Get exactly one trial by immutable trial ID.

docudevs knowledge-base contract-analysis trial-get CASE_ID TRIAL_ID

knowledge-base contract-analysis trial-attest​

Record expert sign-off for an exact immutable trial report.

docudevs knowledge-base contract-analysis trial-attest CASE_ID TRIAL_ID --report-id REPORT_ID
  • --report-id TEXT: Report ID to attest (required)
  • --reason TEXT: Optional reason

knowledge-base contract-analysis run-create​

Create a run and return its immutable run ID.

docudevs knowledge-base contract-analysis run-create KNOWLEDGE_BASE_ID \
  --anchor-job-guid JOB_GUID \
  --idempotency-key run-1
  • --anchor-job-guid TEXT: Anchor job GUID (required)
  • --idempotency-key TEXT: Idempotency key (required)
  • --mode [published|draft_trial]: Run mode (default: published)
  • --draft-revision INTEGER: Required when --mode draft_trial
  • --package-members PATH: JSON array of selected package documents (maximum 32)
  • --expected-missing PATH: JSON array of unavailable expected package documents
  • --excluded-documents PATH: JSON array of explicitly excluded package documents

knowledge-base contract-analysis run-get​

Get status for exactly one immutable run ID.

docudevs knowledge-base contract-analysis run-get RUN_ID

knowledge-base contract-analysis run-result​

Get the result for exactly one immutable run ID.

docudevs knowledge-base contract-analysis run-result RUN_ID

knowledge-base contract-analysis run-reviews​

List the current human review dispositions for an immutable run.

docudevs knowledge-base contract-analysis run-reviews RUN_ID

knowledge-base contract-analysis review​

Append a human disposition to one captured result finding.

docudevs knowledge-base contract-analysis review RUN_ID CRITERION_ID TARGET_KEY \
  --result-fingerprint sha256:... \
  --disposition accepted \
  --expected-review-revision 0 \
  --idempotency-key review-1
  • --result-fingerprint TEXT: Result fingerprint being reviewed (required)
  • --disposition [accepted|corrected|dismissed|unresolved]: Disposition (required)
  • --expected-review-revision INTEGER: Observed review revision (required)
  • --idempotency-key TEXT: Idempotency key (required)
  • --corrected-assessment TEXT: Optional corrected assessment
  • --reason TEXT: Optional reason
  • --evidence PATH: Optional JSON array of immutable evidence references

knowledge-base contract-analysis propose-test-case​

Propose an unapproved suite expectation from one immutable finding.

docudevs knowledge-base contract-analysis propose-test-case CASE_ID \
  --run-id RUN_ID \
  --result-fingerprint sha256:... \
  --criterion-id hours \
  --target-key item:clause-1 \
  --name "Hours threshold" \
  --expected-assessment meets_requirement \
  --reason "Correction" \
  --expected-suite-revision 3 \
  --idempotency-key proposal-1
  • --run-id, --result-fingerprint, --criterion-id, --target-key, --name, --reason, --expected-suite-revision, --idempotency-key: All required
  • --expected-assessment [meets_requirement|deviates_from_requirement|not_found|needs_review|not_applicable]: Required
  • --scenario-tag TEXT: Optional scenario tag (repeatable)

Operations Management​

Run post-processing operations such as error analysis and generative tasks.

operations submit​

Submit an operation by type.

docudevs operations submit JOB_GUID --type error-analysis --parameter quality=deep
  • --type TEXT: Operation type to execute (required)
  • --llm-type [DEFAULT|MINI|HIGH]: Optional LLM override
  • --parameter key=value: Custom operation parameter (repeatable)
  • --wait/--no-wait: Wait for the operation result (default: --no-wait)
  • --timeout INTEGER: Wait timeout in seconds (default: 120)
  • --poll-interval FLOAT: Polling interval in seconds (default: 2.0)

operations contract-analysis​

Run contract analysis against a published knowledge-base profile.

docudevs operations contract-analysis JOB_GUID --knowledge-base-id KNOWLEDGE_BASE_ID
  • --knowledge-base-id INTEGER (alias --knowledge-base): Knowledge base ID (required)
  • --wait/--no-wait: Wait for the result (default: wait)
  • --timeout FLOAT: Wait timeout in seconds (default: 120)
  • --poll-interval FLOAT: Polling interval in seconds (default: 2.0)

operations error-analysis​

Convenience command for error analysis.

docudevs operations error-analysis JOB_GUID --timeout 180
  • --llm-type [DEFAULT|MINI|HIGH]: Optional LLM fallback
  • --parameter key=value: Custom parameter (repeatable)
  • --wait/--no-wait: Wait for completion (default: wait)
  • --timeout INTEGER: Wait timeout in seconds (default: 120)
  • --poll-interval FLOAT: Polling interval in seconds (default: 2.0)

operations generative-task​

Create a generative task from a completed job.

docudevs operations generative-task PARENT_JOB_GUID --prompt "Summarize the findings" --model DEFAULT
  • --prompt TEXT: Prompt for the generative task (required)
  • --model TEXT: Optional LLM model override
  • --temperature FLOAT and --max-tokens INTEGER mirror API parameters
  • --wait/--no-wait: Wait for task completion (default: wait); --no-wait returns immediately with the operation job GUID
  • --timeout INTEGER: Wait timeout in seconds (default: 120)
  • --poll-interval FLOAT: Polling interval in seconds (default: 2.0)

operations pdf-acroform​

Convert a completed PDF job into a generated fillable AcroForm PDF.

docudevs operations pdf-acroform PARENT_JOB_GUID \
  --ocr PREMIUM \
  --page 1 \
  --page 2 \
  --min-confidence 0.35 \
  --max-fields-per-page 300 \
  --timeout 1200 \
  --poll-interval 10 \
  --output medical-examination-form-fillable.pdf
  • Parent job must refer to a PDF
  • --llm-type [DEFAULT|MINI|HIGH]: LLM type for visual field detection (default: DEFAULT)
  • --ocr [PREMIUM|DEFAULT|AUTO]: OCR mode used when page images are missing (default: PREMIUM, the safest choice when the parent job may not already have thumbnails)
  • --page INTEGER: 1-based page number to inspect (repeatable)
  • --min-confidence FLOAT: Minimum accepted field detection confidence
  • --max-fields-per-page INTEGER: Maximum generated fields per page
  • --force-ocr/--no-force-ocr regenerates page images before detection even if thumbnails already exist
  • --wait/--no-wait: Wait for completion (default: wait)
  • --timeout INTEGER / --poll-interval FLOAT: Wait timeout and polling interval (defaults: 180 / 5.0)
  • --output requires the default wait mode because the command downloads the generated PDF after completion
  • The JSON response includes operationJobGuid for later metadata or field-definition retrieval through the SDK or HTTP API

operations status​

List operations created for a job.

docudevs operations status JOB_GUID

operations result​

Fetch the result payload for a completed operation.

docudevs operations result JOB_GUID --type error-analysis

Billing​

The CLI has no billing command group and the Python SDK client has no billing helper methods (there is no list_billing_prices, create_billing_checkout_session, or get_billing_balance). Token packs and balance are managed in the DocuDevs web app; see Billing Tokens.

LLM Provider Management​

Manage LLM providers and key bindings.

warning

llm create and llm update call SDK convenience wrappers that currently send outdated field names and do not match the current provider contract. Use the generated API functions (see Bring Your Own LLM) or the REST API directly instead of these two commands.

llm providers​

List LLM providers.

docudevs llm providers

llm create​

Create an LLM provider.

docudevs llm create --name "My OpenAI" --type OPENAI --api-key "sk-..."

llm get​

Get LLM provider by ID.

docudevs llm get PROVIDER_ID

llm update​

Update LLM provider.

docudevs llm update PROVIDER_ID --name "New Name"

llm delete​

Delete (soft) an LLM provider.

docudevs llm delete PROVIDER_ID

llm keys​

List LLM key bindings.

docudevs llm keys

llm bind​

Bind (or clear) a logical LLM key to a provider.

docudevs llm bind KEY --provider-id PROVIDER_ID

OCR Provider Management​

Manage OCR providers and key bindings.

warning

ocr create and ocr update call SDK convenience wrappers that currently send outdated field names and do not match the current provider contract. Use the generated API functions (see Bring Your Own OCR) or the REST API directly instead of these two commands.

ocr providers​

List OCR providers.

docudevs ocr providers

ocr create​

Create an OCR provider.

docudevs ocr create --name "My Azure OCR" --endpoint "https://..." --api-key "..."

ocr get​

Get OCR provider by ID.

docudevs ocr get PROVIDER_ID

ocr update​

Update OCR provider.

docudevs ocr update PROVIDER_ID --name "New Name"

ocr delete​

Delete (soft) an OCR provider.

docudevs ocr delete PROVIDER_ID

ocr keys​

List OCR key bindings.

docudevs ocr keys

ocr bind​

Bind (or clear) an OCR key to a provider.

docudevs ocr bind KEY --provider-id PROVIDER_ID

Low-Level Commands​

These commands provide direct access to individual API operations.

status​

Check the status of a processing job.

docudevs status JOB_GUID

result​

Get the result of a completed job.

docudevs result JOB_GUID

Case Agent and Evaluation​

agent actions​

List the pending-action inbox of a case.

docudevs agent actions CASE_ID [--status STATUS] [--limit N] [--before ID]

agent confirm​

Confirm a pending action as the signed-in person. The stored request is replayed.

docudevs agent confirm ACTION_ID

agent reject​

Reject a pending action, optionally with a reason.

docudevs agent reject ACTION_ID [--reason TEXT]

agent run start​

Start a headless agent run. PLAYBOOK is general (needs --goal) or investigate-flagged (needs job_id in --params-file). --budget-tokens is authorised when the run starts.

docudevs agent run start CASE_ID investigate-flagged --params-file params.json --budget-tokens 60000 --wait

params.json holds max_rows, flags and job_id.

agent run status​

Show a run's state and report. --wait polls until it finishes.

docudevs agent run status CASE_ID RUN_ID --wait

agent run cancel​

Cancel a run after its current turn.

docudevs agent run cancel CASE_ID RUN_ID

eval holdout​

Start a holdout evaluation. --variants-file is a JSON list of {"name", "memo_include"?, "config"?}; the first entry is the baseline.

docudevs eval holdout CASE_ID --variants-file variants.json --sample-size 200 --seed 7 --group-field base_product_id --wait

eval status​

Show an evaluation, waiting for it with --wait.

docudevs eval status CASE_ID EVAL_ID --wait

eval compare​

Score two completed evaluations against each other. Choose variants with --baseline-variant and --candidate-variant.

docudevs eval compare CASE_ID BASELINE_EVAL_ID CANDIDATE_EVAL_ID --baseline-variant baseline --candidate-variant with-rule

Global Options​

These options are available for all commands:

  • --api-url TEXT: API endpoint URL (default https://api.docudevs.ai)
  • --token TEXT: Authentication token (or use environment variables)
  • --help: Show help message and exit

Output Format​

Most CLI commands return JSON-formatted output that can be parsed by other tools. Commands that produce binary content (for example fill) should be used with --output to write results to disk.

# Save result to file
docudevs process invoice.pdf > result.json

# Parse with jq
docudevs list-configurations | jq '.[] | .name'

# Check if command succeeded
if docudevs process document.pdf; then
    echo "Processing succeeded"
fi

Error Handling​

The CLI returns appropriate exit codes:

  • 0: Success
  • 1: Error (authentication, processing, etc.)

Error messages are written to stderr:

# Redirect errors to file
docudevs process document.pdf 2> errors.log

# Suppress errors
docudevs process document.pdf 2>/dev/null