CLI

All commands run from inside the repository so relative settings, migrations, domain/, and data paths resolve predictably.

Command map

Command	Purpose
sci-rag new	Create a new Sci RAG Kit project.
sci-rag init	Configure an existing Sci RAG Kit checkout in place.
sci-rag doctor	Diagnose the environment: config, domain, database, corpus, credentials.
sci-rag ingest	Ingest documents: parse, chunk, embed, and store them.
sci-rag build	Build the knowledge base in one go: ingest, then extract the graph and its summaries.
sci-rag retrieve	Search the corpus and show the evidence: which layer found each result, and why it ranked.
sci-rag answer	Answer a question from your documents, with numbered citations.
sci-rag stats	Show what is in the knowledge base: documents, chunks, graph size, licenses.
sci-rag profile	Time each retrieval stage (median and 95th percentile) under each profile.
sci-rag serve	Start the web service: REST API under /v1 (docs at /docs) and MCP tools at /mcp.
sci-rag mcp	Let a local agent such as Claude Code use the corpus as tools (MCP over stdio).
sci-rag db	Create or upgrade the database tables.
sci-rag db upgrade	Create or upgrade the database schema (runs the Alembic migrations).
sci-rag graph	Build the knowledge graph: extract concepts and relationships from your documents, then summarize the clusters they form.
sci-rag graph extract	Extract concepts and relationships from ingested chunks.
sci-rag graph communities	Group related entities into clusters and write a summary of each (rebuilds all).
sci-rag graph citations	Link corpus documents that cite each other, from the reference lists corpus enrich fetched.
sci-rag graph resolve-entities	Resolve duplicate graph entities conservatively and audit every merge.
sci-rag graph gc	Remove graph leftovers: entities with no evidence, relationships with a missing end, and communities whose members are gone.
sci-rag campaign	Find papers to add: search by topic or DOI list, check each one's rights, and download the open-access PDFs into a manifest you can ingest.
sci-rag campaign discover	Discover a deduplicated DOI list and save resumable state.
sci-rag campaign build	Check each candidate's rights, download the open-access PDFs, and write a manifest to ingest.
sci-rag campaign screen	Screen discovered abstracts and queue uncertain rows for human review.
sci-rag campaign review	Walk pending screening rows and append explicit human decisions.
sci-rag manifest	Check a corpus manifest (data/corpus.jsonl) before you ingest it.
sci-rag manifest lint	Check a corpus manifest before ingesting it.
sci-rag draft	Draft the domain files you would otherwise hand-write.
sci-rag draft questions	Draft seed questions grounded in your own documents, and verify them.
sci-rag draft seed-from-answers	Draft seed rows for questions you already have, from the kit's own answers.
sci-rag draft manifest	Read title, authors, year, and source off your documents.
sci-rag draft ontology	Redraft or refine the ontology against what your documents actually say.
sci-rag draft prompts	Reword a prompt for your field.
sci-rag eval	Measure quality against your seed questions: retrieval scores, what each retrieval layer contributes, and graded answers.
sci-rag eval retrieval	Score retrieval against your seed questions: did the right documents come back?
sci-rag eval answers	Answer every seed question and grade each answer for grounding, citations, and correctness.
sci-rag eval diff	Compare two eval runs: per-question rank moves and paired metric deltas.
sci-rag eval html	Render an eval run as one self-contained HTML page.
sci-rag eval calibrate	Compare a person's scores with the grader's, per dimension (reported as Cohen's kappa).
sci-rag embed	Re-embed rows left behind after an embedding model change.
sci-rag embed reindex	Re-embed chunks and community summaries that an older embedding model produced.
sci-rag corpus	Look after the corpus: rights report, publication metadata, snapshots, export, and deletion.
sci-rag corpus enrich	Add Crossref citation, journal, and retraction metadata to the corpus.
sci-rag corpus delete	Delete documents, their chunks, and every graph entry that relied on them.
sci-rag corpus snapshot	Record exactly which documents the corpus holds right now, under data/snapshots/.
sci-rag corpus license-report	Count documents and chunks by license class, and name the ones still unknown.
sci-rag corpus export	Export documents, chunks, entities, and relationships to files.

sci-rag new

Create a new Sci RAG Kit project.

$ sci-rag new [OPTIONS]

Options

Option	Type	Default	Description
--output-dir, -o	path	.	Where to create the project directory. Defaults to the current directory.
--defaults	boolean	false	Take every default without asking. Useful in CI.
--answers-file	path	unset	A YAML file of answers, for reproducible generation. Unanswered questions take their default. Answers that need a person, such as accepting a drafted ontology, must be supplied explicitly.
--quick, --advanced	boolean	unset	Ask six setup questions, or expose every option.
--no-tty	boolean	false	Use plain numbered prompts even in a supported terminal.
--no-preflight	boolean	false	Skip the live credential check before downloading the template.
--ref	text	unset	Fetch the template at this tag or branch, overriding the tag that matches this generator's version.
--template-path	path	unset	Generate from a local checkout with no network access. Copies what the checkout tracks, so local state stays local.

sci-rag init

Configure an existing Sci RAG Kit checkout in place. The setup wizard asks about your project, LLM provider credentials, domain, corpus, and stack, then updates the target checkout's configuration files. It writes editable configuration files.

$ sci-rag init [OPTIONS]

Options

Option	Type	Default	Description
--target	path	.	The checkout to configure. Defaults to the current directory.
--defaults	boolean	false	Take every default without asking. Useful in CI.
--answers-file	path	unset	A YAML file of answers, for reproducible generation. Unanswered questions take their default. Answers that need a person, such as accepting a drafted ontology, must be supplied explicitly.
--quick, --advanced	boolean	unset	Ask six setup questions, or expose every option.
--no-tty	boolean	false	Use plain numbered prompts even in a supported terminal.
--dry-run	boolean	false	Show what would change without writing anything.

sci-rag doctor

Diagnose the environment: config, domain, database, corpus, credentials.

$ sci-rag doctor [OPTIONS]

Options

Option	Type	Default	Description
--probe	boolean	false	Also make one tiny live embedding and generation call.

sci-rag ingest

Ingest documents: parse, chunk, embed, and store them. Pass a folder for a first run (every document is recorded with license_class "unknown"), or a manifest once you have declared each document's rights. sci-rag build runs this and the graph steps together.

$ sci-rag ingest [OPTIONS] [PATH]

Arguments

Name	Type	Default	Description
PATH	path	unset	Folder of documents to ingest (PDF, HTML, Markdown, or plain text).

Options

Option	Type	Default	Description
--manifest, -m	path	unset	A JSONL corpus manifest: one line per document with title, authors, and rights.
--source	text	local	Source label recorded on documents ingested from a folder.
--no-docling	boolean	false	Skip Docling even if installed (use the pypdf fallback).
--chunk-tokens	integer	800	Target tokens per chunk.
--overlap-tokens	integer	150	Overlap tokens between chunks.

sci-rag build

Build the knowledge base in one go: ingest, then extract the graph and its summaries. Ingestion works with any embedding setup, including offline. The graph steps need LLM provider credentials; without them, or with --no-graph, they are skipped and the command says so. Re-running only processes new chunks.

$ sci-rag build [OPTIONS] [PATH]

Arguments

Name	Type	Default	Description
PATH	path	unset	Folder of documents to ingest (PDF, HTML, Markdown, or plain text).

Options

Option	Type	Default	Description
--manifest, -m	path	unset	A JSONL corpus manifest: one line per document with title, authors, and rights.
--source	text	local	Source label recorded on documents ingested from a folder.
--no-docling	boolean	false	Skip Docling even if installed (use the pypdf fallback).
--no-graph	boolean	false	Stop after ingestion; skip the knowledge graph.
--batch-size	integer	10	Chunks per graph-extraction model call.

sci-rag retrieve

Search the corpus and show the evidence: which layer found each result, and why it ranked.

$ sci-rag retrieve [OPTIONS] QUERY

Arguments

Name	Type	Default	Description
QUERY	text	required	The question to search for.

Options

Option	Type	Default	Description
--profile	text	deep	Which retrieval layers run: interactive (vector and keyword, fast), deep (all five), or auto (a router picks per question).
--limit	integer	8	How many results to return after the layers are combined.
--license	text	unset	Comma-separated license allowlist (e.g. public,open_commercial).
--source	text	unset	Comma-separated source allowlist.
--year-min	integer	unset	Earliest publication year to include.
--year-max	integer	unset	Latest publication year to include.
--author	text	unset	Comma-separated author allowlist (exact strings).
--journal	text	unset	Comma-separated journal allowlist.
--exclude-doi	text	unset	Comma-separated DOIs to drop.
--explain-routing	boolean	false	Print what the auto router decides for this query (and why) before retrieving.

sci-rag answer

Answer a question from your documents, with numbered citations. Needs LLM provider credentials.

$ sci-rag answer [OPTIONS] QUERY

Arguments

Name	Type	Default	Description
QUERY	text	required	The question to answer.

Options

Option	Type	Default	Description
--profile	text	deep	Which retrieval layers run: interactive (vector and keyword, fast), deep (all five), or auto (a router picks per question).
--limit	integer	8	How many sources to give the model.
--license	text	unset	Comma-separated license allowlist (e.g. public,open_commercial).
--source	text	unset	Comma-separated source allowlist.
--year-min	integer	unset	Earliest publication year to include.
--year-max	integer	unset	Latest publication year to include.
--author	text	unset	Comma-separated author allowlist (exact strings).
--journal	text	unset	Comma-separated journal allowlist.
--exclude-doi	text	unset	Comma-separated DOIs to drop.
--include-retracted	boolean	false	Deliberately allow known retracted papers as answer evidence.
--compression, --no-compression	boolean	unset	Summarize each retrieved source before answering, or send the full text. Omit to use the domain profile's setting.

sci-rag stats

Show what is in the knowledge base: documents, chunks, graph size, licenses.

$ sci-rag stats [OPTIONS]

sci-rag profile

Time each retrieval stage (median and 95th percentile) under each profile. Replays the seed questions against interactive, deep, and auto, and aggregates the per-stage timings every request already records. Stages run concurrently, so their durations do not sum to the request; wall-clock is measured separately and reported beside them.

$ sci-rag profile [OPTIONS]

Options

Option	Type	Default	Description
--questions	path	unset	Seed questions JSONL. Defaults to the domain's.
--runs	integer	3	Replays per question, per profile.
--limit	integer	8	Results per request.
--json	boolean	false	Machine-readable output.

sci-rag serve

Start the web service: REST API under /v1 (docs at /docs) and MCP tools at /mcp.

$ sci-rag serve [OPTIONS]

Options

Option	Type	Default	Description
--host	text	unset	Bind address (default from settings).
--port	integer	unset	Port (default from settings; Cloud Run sets PORT).

sci-rag mcp

Let a local agent such as Claude Code use the corpus as tools (MCP over stdio). Add it to an agent with, for example: claude mcp add sci-rag -- uv run --directory /path/to/your/repo sci-rag mcp

$ sci-rag mcp [OPTIONS]

sci-rag db

Create or upgrade the database tables.

$ sci-rag db [OPTIONS] COMMAND [ARGS]...

sci-rag db upgrade

Create or upgrade the database schema (runs the Alembic migrations).

$ sci-rag db upgrade [OPTIONS]

sci-rag graph

Build the knowledge graph: extract concepts and relationships from your documents, then summarize the clusters they form.

$ sci-rag graph [OPTIONS] COMMAND [ARGS]...

sci-rag graph extract

Extract concepts and relationships from ingested chunks. Needs LLM provider credentials. Only chunks the extractor has not seen are processed, so re-running after a new ingest picks up just the new documents.

$ sci-rag graph extract [OPTIONS]

Options

Option	Type	Default	Description
--batch-size	integer	10	Chunks per extraction call.
--all	boolean	false	Re-read every chunk, including previously processed chunks.
--max-chunks	integer	unset	Stop after this many chunks (for trials).

sci-rag graph communities

Group related entities into clusters and write a summary of each (rebuilds all). Summaries answer big-picture questions no single passage covers. Needs a LLM provider credentials.

$ sci-rag graph communities [OPTIONS]

Options

Option	Type	Default	Description
--min-size	integer	3	Smallest cluster worth summarizing.

sci-rag graph citations

Link corpus documents that cite each other, from the reference lists corpus enrich fetched.

$ sci-rag graph citations [OPTIONS]

Options

Option	Type	Default	Description
--dry-run, --apply	boolean	true	Preview by default; --apply reconciles cached Crossref references.

sci-rag graph resolve-entities

Resolve duplicate graph entities conservatively and audit every merge.

$ sci-rag graph resolve-entities [OPTIONS]

Options

Option	Type	Default	Description
--dry-run, --apply	boolean	true	Preview merges by default; --apply merges duplicates and records each merge in an audit table.
--no-llm	boolean	false	Skip borderline pairs. Nothing is sent to a model.
--threshold	float range	0.92	Minimum same-type similarity for an automatic fuzzy merge.
--llm-threshold	float range	0.8	Minimum similarity for a borderline pair to be reviewed by the LLM.

sci-rag graph gc

Remove graph leftovers: entities with no evidence, relationships with a missing end, and communities whose members are gone.

$ sci-rag graph gc [OPTIONS]

Options

Option	Type	Default	Description
--dry-run, --apply	boolean	true	--dry-run (default) reports what would go; --apply removes it.

sci-rag campaign

Find papers to add: search by topic or DOI list, check each one's rights, and download the open-access PDFs into a manifest you can ingest.

$ sci-rag campaign [OPTIONS] COMMAND [ARGS]...

sci-rag campaign discover

Discover a deduplicated DOI list and save resumable state.

$ sci-rag campaign discover [OPTIONS]

Options

Option	Type	Default	Description
--topic	text	unset	Search topic for OpenAlex discovery.
--doi-file	file	unset	Text file with one DOI or DOI URL per line.
--name	text	unset	Campaign directory name (derived from the input when omitted).
--mailto	text	required	Contact email sent to OpenAlex and Crossref.
--max-results	integer range	100	Maximum total candidates for a topic campaign.
--campaign-root	directory	data/campaigns	Parent directory for campaign state.

sci-rag campaign build

Check each candidate's rights, download the open-access PDFs, and write a manifest to ingest.

$ sci-rag campaign build [OPTIONS]

Options

Option	Type	Default	Description
--topic	text	unset	Search topic for OpenAlex discovery.
--doi-file	file	unset	Text file with one DOI or DOI URL per line.
--name	text	unset	Campaign directory name (derived from the input when omitted).
--mailto	text	required	Contact email sent to OpenAlex, Crossref, and Unpaywall.
--dry-run	boolean	false	Resolve and report rights without downloading PDFs or writing a manifest.
--max-results	integer range	100	Work on at most this many candidates, including ones already in state.
--all-candidates	boolean	false	Ignore --max-results and process every candidate the campaign retains.
--max-pdf-mb	integer range	25	Reject a PDF larger than this many MiB.
--campaign-root	directory	data/campaigns	Parent directory for campaign state, PDFs, and manifest.

sci-rag campaign screen

Screen discovered abstracts and queue uncertain rows for human review.

$ sci-rag campaign screen [OPTIONS]

Options

Option	Type	Default	Description
--name	text	required	Campaign directory name to screen.
--criteria-file	file	required	Plain-text inclusion and exclusion criteria.
--confidence-threshold	float range	0.8	Model confidence below this value requires human review.
--batch-size	integer range	20	Maximum abstracts in one model request.
--campaign-root	directory	data/campaigns	Parent directory containing campaign state.

sci-rag campaign review

Walk pending screening rows and append explicit human decisions.

$ sci-rag campaign review [OPTIONS]

Options

Option	Type	Default	Description
--name	text	required	Campaign directory name to review.
--campaign-root	directory	data/campaigns	Parent directory containing campaign state.

sci-rag manifest

Check a corpus manifest (data/corpus.jsonl) before you ingest it.

$ sci-rag manifest [OPTIONS] COMMAND [ARGS]...

sci-rag manifest lint

Check a corpus manifest before ingesting it. Ingestion reports a bad manifest one document at a time, after the run has already started. This reports every problem at once, before anything is parsed, embedded, or written.

$ sci-rag manifest lint [OPTIONS] PATH

Arguments

Name	Type	Default	Description
PATH	path	required	The JSONL corpus manifest to check.

sci-rag draft

Draft the domain files you would otherwise hand-write. Every drafter can also print its corpus-bearing prompt for local use or an approved assistant (--print-prompt), then read the reply back (--from-file).

$ sci-rag draft [OPTIONS] COMMAND [ARGS]...

sci-rag draft questions

Draft seed questions grounded in your own documents, and verify them.

$ sci-rag draft questions [OPTIONS]

Options

Option	Type	Default	Description
--count	integer	10	How many questions to ask for.
--folder	path	unset	Draft from documents in this folder, before anything is ingested.
--print-prompt	boolean	false	Print the rendered prompt and exit. It may contain corpus text; send it only to an approved destination.
--from-file	path	unset	Read the model's reply from this file, with no model call.
--output	path	unset	Where to write the proposal. Defaults to .proposed.
--apply	boolean	false	Append the verified questions to the seed file.
--dry-run	boolean	false	Show what would be drafted without writing anything.
--repair, --no-repair	boolean	true	Ask the model once more to replace rows that failed grounding verification.

sci-rag draft seed-from-answers

Draft seed rows for questions you already have, from the kit's own answers. draft questions invents the questions. This takes yours, answers each one, and proposes ground truth from the evidence that answer cited. Nothing is taken on the model's word: evidence phrases are extracted from the retrieved chunk text and every row is checked against the same relevance predicate the evaluation uses, so a row that would score zero against its own evidence is dropped and never proposed.

$ sci-rag draft seed-from-answers [OPTIONS] QUESTIONS_FILE

Arguments

Name	Type	Default	Description
QUESTIONS_FILE	path	required	Plain text file of questions, one per line. # comments are skipped.

Options

Option	Type	Default	Description
--profile	text	deep	Retrieval profile to answer with.
--limit	integer	8	Sources per answer.
--output	path	unset	Where to write the proposal. Defaults to .proposed.
--apply	boolean	false	Append the proposed rows to the seed file.
--dry-run	boolean	false	Show what would be proposed without writing anything.

sci-rag draft manifest

Read title, authors, year, and source off your documents. Rights stay yours.

$ sci-rag draft manifest [OPTIONS]

Options

Option	Type	Default	Description
--folder	path	unset	Documents to describe. Defaults to /raw.
--print-prompt	boolean	false	Print the rendered prompt and exit. It may contain corpus text; send it only to an approved destination.
--from-file	path	unset	Read the model's reply from this file, with no model call.
--output	path	unset	Where to write the proposal. Defaults to .proposed.
--manifest	path	unset	The manifest being drafted. Defaults to /corpus.jsonl.
--apply	boolean	false	Write the manifest directly, with no .proposed file to review.
--dry-run	boolean	false	Show what would be drafted without writing anything.
--batch-size	integer	unset	Documents per model call. Defaults to 12.

sci-rag draft ontology

Redraft or refine the ontology against what your documents actually say.

$ sci-rag draft ontology [OPTIONS]

Options

Option	Type	Default	Description
--from-corpus	boolean	false	Draft from passages in your ingested corpus (or data/raw before ingestion). This is the default; the flag names it explicitly.
--refine	boolean	false	Show the model your ontology and ask only what it would add and remove.
--cold	boolean	false	Draft from the description alone, without reading any document.
--folder	path	unset	Draft from documents in this folder, before anything is ingested.
--print-prompt	boolean	false	Print the rendered prompt and exit. It may contain corpus text; send it only to an approved destination.
--from-file	path	unset	Read the model's reply from this file, with no model call.
--output	path	unset	Where to write the proposal. Defaults to .proposed.
--apply	boolean	false	Write domain.yaml directly, with no .proposed file to review.
--dry-run	boolean	false	Show what would change without writing anything.

sci-rag draft prompts

Reword a prompt for your field. Judge prompts are refused by name.

$ sci-rag draft prompts [OPTIONS] NAME

Arguments

Name	Type	Default	Description
NAME	text	required	Which prompt to reword: entity_extraction or answer.

Options

Option	Type	Default	Description
--print-prompt	boolean	false	Print the rendered prompt and exit. It may contain corpus text; send it only to an approved destination.
--from-file	path	unset	Read the model's reply from this file, with no model call.
--output	path	unset	Where to write the proposal. Defaults to .md.proposed.
--apply	boolean	false	Write the prompt file directly, with no .proposed file to review.
--dry-run	boolean	false	Show the rewrite without writing anything.

sci-rag eval

Measure quality against your seed questions: retrieval scores, what each retrieval layer contributes, and graded answers.

$ sci-rag eval [OPTIONS] COMMAND [ARGS]...

sci-rag eval retrieval

Score retrieval against your seed questions: did the right documents come back?

$ sci-rag eval retrieval [OPTIONS]

Options

Option	Type	Default	Description
--questions	path	unset	Seed questions JSONL.
--limit	integer	10	Results retrieved per question.
--ablation	boolean	false	Also score each retrieval layer switched off in turn, to see what each contributes.
--condition	text	unset	Label an established corpus condition (currently: resolved_entities).
--snapshot	text	unset	Record this corpus snapshot name in the report.

sci-rag eval answers

Answer every seed question and grade each answer for grounding, citations, and correctness. Needs LLM provider credentials. The grader scores grounding without seeing the reference answer, then scores correctness against it in a separate pass.

$ sci-rag eval answers [OPTIONS]

Options

Option	Type	Default	Description
--questions	path	unset	Seed questions JSONL.
--profile	text	deep	Retrieval profile for answer generation.
--limit	integer	8	Sources per answer.
--judge-model	text	unset	Judge model spec, 'model' or 'provider:model'. Overrides SCI_RAG_JUDGE_MODEL.
--snapshot	text	unset	Record this corpus snapshot name in the report.
--compressed	boolean	false	Summarize each retrieved source before answering, as a separate condition to compare.

sci-rag eval diff

Compare two eval runs: per-question rank moves and paired metric deltas. Deltas are B minus A. Run it after any retrieval-affecting change to see whether the improvement is real or inside the noise.

$ sci-rag eval diff [OPTIONS] REPORT_A REPORT_B

Arguments

Name	Type	Default	Description
REPORT_A	path	required	Baseline report.json (or its run directory).
REPORT_B	path	required	Comparison report.json (or its run directory).

Options

Option	Type	Default	Description
--config	text	unset	Diff only this ablation config (default: every common config).
--output	path	unset	Also write the markdown diff to this path.

sci-rag eval html

Render an eval run as one self-contained HTML page. For sharing with a collaborator who will never open a terminal. Inline styles, nothing fetched when the page is opened, and the small-sample and drafted-ground-truth warnings travel with the numbers. Picks up calibration.json automatically when it sits beside the report.

$ sci-rag eval html [OPTIONS] RUN

Arguments

Name	Type	Default	Description
RUN	path	required	Run directory, or the report.json inside one.

Options

Option	Type	Default	Description
--output	path	unset	Where to write the page (default: report.html beside the report).

sci-rag eval calibrate

Compare a person's scores with the grader's, per dimension (reported as Cohen's kappa). Appends a calibration section to the report's markdown (report.md) and writes calibration.json next to it, so the kappa travels with the eval numbers it qualifies.

$ sci-rag eval calibrate [OPTIONS]

Options

Option	Type	Default	Description
--labels	path	required	Human labels (labels.jsonl).
--report	path	unset	Answers report.json (or its run directory) to calibrate against. Defaults to the newest answers run under eval_results/.
--output	path	unset	Also write the calibration markdown to this path.

sci-rag embed

Re-embed rows left behind after an embedding model change.

$ sci-rag embed [OPTIONS] COMMAND [ARGS]...

sci-rag embed reindex

Re-embed chunks and community summaries that an older embedding model produced. A dimension change is refused outright: that is a schema migration plus a full re-ingest, never a reindex.

$ sci-rag embed reindex [OPTIONS]

Options

Option	Type	Default	Description
--dry-run, --apply	boolean	true	--dry-run (default) reports what is stale; --apply re-embeds it.
--batch-size	integer	32	Rows re-embedded per batch (one commit per batch).

sci-rag corpus

Look after the corpus: rights report, publication metadata, snapshots, export, and deletion.

$ sci-rag corpus [OPTIONS] COMMAND [ARGS]...

sci-rag corpus enrich

Add Crossref citation, journal, and retraction metadata to the corpus.

$ sci-rag corpus enrich [OPTIONS]

Options

Option	Type	Default	Description
--mailto	text	required	Contact email sent with each Crossref request; identified callers get faster service.
--dry-run	boolean	false	List eligible DOI records without network calls or writes.
--limit	integer range	unset	Process at most this many documents.

sci-rag corpus delete

Delete documents, their chunks, and every graph entry that relied on them. Communities built from that evidence are dropped too; rebuild them with sci-rag graph communities. Run sci-rag graph gc afterwards to remove entities left with no evidence at all.

$ sci-rag corpus delete [OPTIONS] DOCUMENT_IDS...

Arguments

Name	Type	Default	Description
DOCUMENT_IDS	text, repeatable	required	Document id(s) to delete.

Options

Option	Type	Default	Description
--yes, -y	boolean	false	Skip the confirmation prompt.

sci-rag corpus snapshot

Record exactly which documents the corpus holds right now, under data/snapshots/. Records counts, per-document content hashes, embedding versions, the git commit, and a single corpus digest. Reference it from eval runs with --snapshot NAME so reported numbers stay tied to exactly the corpus that produced them.

$ sci-rag corpus snapshot [OPTIONS] [NAME]

Arguments

Name	Type	Default	Description
NAME	text	unset	Snapshot name (default: UTC timestamp).

sci-rag corpus license-report

Count documents and chunks by license class, and name the ones still unknown. License classes decide what a scoped request may see, so this is the report to read before you share the service.

$ sci-rag corpus license-report [OPTIONS]

Options

Option	Type	Default	Description
--json	boolean	false	Machine-readable output.
--strict	boolean	false	Exit 1 when any document is still 'unknown'. For CI; the report itself never fails.

sci-rag corpus export

Export documents, chunks, entities, and relationships to files. A scope leaves out anything it cannot vouch for, and applies to the graph too: an entity is exported only when every document it was extracted from is in scope, because its description is written from all of them. Communities are never exported; they aggregate with no per-document attribution to check.

$ sci-rag corpus export [OPTIONS] OUTDIR

Arguments

Name	Type	Default	Description
OUTDIR	path	required	Directory to write the export files into.

Options

Option	Type	Default	Description
--format	text	jsonl	jsonl (no extra dependency) or parquet (needs --extra export).
--license	text	empty	Export only these license classes (repeatable). Omit for everything. An export is a redistribution, so a scope excludes 'unknown' unless you name it.
--include-embeddings	boolean	false	Include chunk vectors (large, rarely wanted).

Shell help is authoritative too

Run uv run sci-rag COMMAND --help for the same current command tree in the terminal. The generated page is a browsing aid; Click remains the parser that decides whether a concrete invocation is valid.