api.spokenalpha.com
Detect when CEOs and CFOs deviate from their own language baseline — the leading indicator of forward-print disappointment. Structured JSON. Self-serve signup. 80K+ calls indexed.
CFO · hedge density z-score
Language deviation surfaces before the guidance cut.
What you get
Start with metadata and summaries on the free tier. Unlock transcripts, themes, and per-speaker deviation as you scale.
AI-generated TL;DR per call, available minutes after the transcript drops. Covers prepared remarks and Q&A highlights. Structured JSON — no text parsing required.
AAPL Q1 FY2024: Revenue $119.6B, slightly above consensus. Cook highlighted services momentum; CFO struck cautious tone on China FX headwinds.
See docs →15 cross-company themes — data center capex, GLP-1 exposure, supply chain normalization — surfaced across every public earnings call. Filter by theme, date, or ticker.
data_center_capex · 80K+ calls indexed · updated live
See docs →Full-text search across 80K+ calls. Retrieve parsed Q&A exchanges, statement counts, and speaker attribution. Results return in JSON with call metadata.
GET /v1/transcripts/{call_id} · Q&A pairs with role attribution
See docs →The advanced tier. Per-speaker longitudinal baseline tracks how each executive's language changes across quarters. Includes hedge density, qualifier load, and behavioral deviation scoring.
Each speaker measured against their own call history — not a generic sentiment index.
See pricing →Want to test the per-speaker scoring logic without building an integration? Paste any text and see how it scores against a speaker's historical baseline.
Try it live →Why Spoken Alpha
A CFO who suddenly hedges 47% more than they ever have isn't a sentiment reading — it's a tell. We measure each executive against their own call history. Marc Casper (TMO) sounds different from every other CEO; his baseline is his own.
The signal is strongest at the extremes. High-conviction deviation calls — where multiple speakers show correlated breaks from their baselines on the same call — are where the data earns its cost.
1,000 API calls per month, no card required. Earnings metadata and call lists are free tier. Upgrade when you're ready to move to production. No enterprise negotiation needed under $500/mo.
Try it now
The /v1/calls/{call_id}/executive-summary endpoint returns a structured summary with key themes and tone. Available on Starter and above.
curl -H "Authorization: Bearer sk_live_..." \ https://api.spokenalpha.com/v1/calls/42891/executive-summary
{
"call_id": "42891",
"ticker": "AAPL",
"company_name": "Apple Inc.",
"quarter": "Q1 FY2024",
"call_date": "2024-02-01",
"summary": "Apple reported Q1 FY2024 revenue of $119.6B, slightly above consensus. Cook highlighted services momentum and iPhone resilience. CFO Maestri struck a notably cautious tone on China macro, citing 'meaningful FX headwinds' — a shift in hedging language relative to prior calls.",
"speakers_covered": 4,
"duration_minutes": 51,
"key_themes": [
"services growth",
"China macro headwinds",
"FX impact"
],
"generated_at": "2024-02-01T23:14:29Z"
}Pro endpoint
The POST /v1/text/score endpoint accepts any passage — earnings call excerpt, press release, CNBC interview transcript — and returns z-scores measuring how much the language deviates from that speaker's own historical baseline. No call ID required; paste any text.
import anthropic_spoken_alpha as sa
client = sa.Client(api_key="sk_live_...")
result = client.text.score(
text="[Paste the passage you want to score here — "
"earnings call excerpt, press release, or interview transcript]",
speaker="Jane Doe", # swap in the executive's name
company="ACME", # optional — disambiguates common names
context={"segment_type": "qa_answer"},
)
print(result.scores.hedge_density.z) # e.g. +2.3
print(result.aggregate.dominant_dimension) # e.g. "hedge_density"Example response (Pro tier)
{
"speaker_resolved": {
"display_name": "Jane Doe",
"role": "CFO",
"company": "ACME",
"baseline_call_count": 14
},
"scores": {
"hedge_density": { "raw": 0.041, "z": 2.3, "percentile": 89 },
"forward_looking_ratio": { "raw": 0.18, "z": -1.5, "percentile": 12 },
"evasion_proxy": null,
"tokens_per_turn": { "raw": 87.0, "z": -0.4, "percentile": 34 }
},
"aggregate": {
"max_abs_z": 2.3,
"dominant_dimension": "hedge_density",
"tokens_scored": 312
}
}Pricing
Free tier requires no card. Paid tiers unlock transcripts, summaries, and per-speaker deviation. Overage-metered, not hard-capped on paid plans.
Evaluate and prototype. Metadata endpoints only.
Production agents and research workflows.
Full access including per-speaker deviation.
Enterprise
Custom quotas, SSO, dedicated SLAs, on-prem options. Volume pricing negotiated per contract.
Start free — no card, no enterprise negotiation. Upgrade when your data needs grow.
Built for buy-side analysts, fund managers, and fintech developers.
© 2026 Spoken Alpha Inc. · api.spokenalpha.com