What inputs does save_claim accept?

statement (string, required): The atomic, falsifiable statement. One claim, not a paragraph. tickers (array, required): Entities referenced (uppercased). 1 for most claims; 2+ for a comparison claim. claim_type ("assertion" | "prediction" | "judgment", required): Epistemic type — drives scoring. assertion=true now (verified vs data); prediction=future (resolves at horizon via verifiable_condition); judgment=qualitative (not auto-scored). tags (array, optional, default []): Topical labels (controlled vocab). Multi-valued; drives filtering + learning segmentation, not scoring. direction ("bullish" | "bearish" | "neutral", required): Directional polarity of the claim. confidence (number, required): Author confidence in [0,1]. Used as the Brier/log-loss weight when scored. eval_mode ("auto" | "agent" | "manual", optional, default "agent"): How the outcome is resolved: auto (deterministic grade vs data via verifiable_condition), agent (an LLM judges at resolution), manual (a human marks it). verifiable_condition (object|null, optional, default null): Machine-evaluable condition for eval_mode='auto'. Null otherwise. antecedent (string|null, optional, default null): Scenario precondition — the claim only resolves when this holds. Null = unconditional. horizon_days (integer|null, optional, default null): Resolution horizon in days (predictions). Null for assertions/judgments. evidence (object, optional, default {"fact_ids":[],"source_urls":[]}): Evidence grounding the claim. source_report_id (string, optional): Optional id of a report that contains the supporting analysis. visibility ("private" | "unlisted" | "public", optional, default "private"): 'private' (default) owner-only; 'unlisted' visible at a direct URL; 'public' surfaces on the author's profile and contributes to the claim-accuracy reputation. idempotency_key (string, optional): Optional client key for at-most-once semantics from a retrying agent.

Which Valuein plans include save_claim?

Available on the free tier (token required) and on Pro and Institutional. Not point-in-time safe.

Discovery

Save Claim

save_claim

Benchmark and above

Persist a single falsifiable, evidence-backed CLAIM — the atomic unit of the research graph. Use this for each discrete assertion an analysis produces (e.g. 'NVDA gross margin stays above 70% through FY2026'), then compose claims into a thesis with `link_claim_to_thesis`. Clai...

Example Call

example_call.pypython

# Using the Valuein MCP server from Python (via MCP SDK)# Or call directly from Claude / Cursor after setup result = await client.call_tool(    "save_claim",    arguments={    "statement": "EXAMPLE",    "tickers": [],    "claim_type": "assertion",    "direction": "bullish",    "confidence": 10})print(result)

Direct tool call:save_claim(...)

Inputs

Parameter	Type	Required	Description
`statement`	string	required	The atomic, falsifiable statement. One claim, not a paragraph.
`tickers`	array	required	Entities referenced (uppercased). 1 for most claims; 2+ for a comparison claim.
`claim_type`	"assertion" \| "prediction" \| "judgment"	required	Epistemic type — drives scoring. assertion=true now (verified vs data); prediction=future (resolves at horizon via verifiable_condition); judgment=qualitative (not auto-scored).
`tags`	array	optional	Topical labels (controlled vocab). Multi-valued; drives filtering + learning segmentation, not scoring.Default: `[]`
`direction`	"bullish" \| "bearish" \| "neutral"	required	Directional polarity of the claim.
`confidence`	number	required	Author confidence in [0,1]. Used as the Brier/log-loss weight when scored.
`eval_mode`	"auto" \| "agent" \| "manual"	optional	How the outcome is resolved: auto (deterministic grade vs data via verifiable_condition), agent (an LLM judges at resolution), manual (a human marks it).Default: `"agent"`
`verifiable_condition`	object\|null	optional	Machine-evaluable condition for eval_mode='auto'. Null otherwise.Default: `null`
`antecedent`	string\|null	optional	Scenario precondition — the claim only resolves when this holds. Null = unconditional.Default: `null`
`horizon_days`	integer\|null	optional	Resolution horizon in days (predictions). Null for assertions/judgments.Default: `null`
`evidence`	object	optional	Evidence grounding the claim.Default: `{"fact_ids":[],"source_urls":[]}`
`source_report_id`	string	optional	Optional id of a report that contains the supporting analysis.
`visibility`	"private" \| "unlisted" \| "public"	optional	'private' (default) owner-only; 'unlisted' visible at a direct URL; 'public' surfaces on the author's profile and contributes to the claim-accuracy reputation.Default: `"private"`
`idempotency_key`	string	optional	Optional client key for at-most-once semantics from a retrying agent.

Output Fields

claimdeduplicatedcapacity

Example Response

Illustrative

response.jsonjson

{  "claim": null,  "deduplicated": null,  "capacity": null}

Shape only — field names match the live schema; values are placeholders, not real filings. See MCP server for the structured response shape.

Notes

Available from the Benchmark plan and above.

Try this tool

Get token All tools

CategoryDiscovery

PIT-safeNo

Min planBenchmark

Inputs14 (5 required)

Plan Access

Sample (guest)

Not included

Benchmark (token)

Available

Pro

Available

Institutional

Available

Related Tools

search_companies

Search Companies

get_company_fundamentals

Company Fundamentals

get_valuation_metrics

Valuation Metrics

get_financial_ratios

Financial Ratios

Resources

Point-in-Time methodology →Parquet schema reference →Bulk Data API reference →Python SDK guide →

Index names are trademarks of their owners; Valuein is not affiliated with or endorsed by any index provider.