Hi team — I’m building a stock analysis agent and would love to use data from the Perplexity Finance pages (e.g. https://www.perplexity.ai/finance/FICO). Those pages contain uniquely valuable data that I can’t find elsewhere, in particular:
- Daily narrative timeline — for each trading day: close, change%, and a 1-2 paragraph English summary synthesizing that day’s catalysts (regulatory, analyst actions, earnings, macro). This is Perplexity’s signature output and massively useful for building a catalyst map.
- Full analyst action history — each firm’s target price changes with dates (e.g. “JPM cut target $1,825 → $1,325 on 3/24”), not just current consensus.
- Real-time quote block — current price, pre/post-market, PE (TTM), market cap, 52w range, day high/low, EPS, etc.
- Earnings analysis with highlights — beat/miss with analyst-style bullets.
What I’ve tried
1. sonar-pro / sonar-reasoning-pro with search_domain_filter: ["perplexity.ai"]
The model’s citations correctly point to perplexity.ai/finance/FICO, but only short snippets come back. When forced into a JSON schema response, the model hallucinates numbers to fill the schema (e.g. returned price_current: 909 when actual was 970.17). The reasoning trace even admits “I can’t fetch
URLs or access real-time data.”
2. Pro Search with fetch_url_content tool
The Pro Search Tools doc mentions a fetch_url_content tool that “retrieves complete content from specified URLs”. In practice, pointing the model at the Perplexity Finance URL does not actually fetch it — the tool appears not to trigger or returns empty
content for these internal pages.
3. /search API with search_domain_filter: ["perplexity.ai"]
Returns results but the content field is only ~60 characters for Perplexity Finance pages (likely because the SPA content isn’t fully indexed). Most returned results are unrelated changelog pages.
4. Browser automation (works, but heavy)
I can reliably extract the data by driving a real Chrome session via CDP, passing Cloudflare Turnstile, then reading window.__NEXT_DATA__ or the internal /rest/finance/* endpoints with session cookies. But this feels like it’s going against the grain — the data is already produced by Perplexity, I’d much
rather pay for API access.
What I’m asking
- Is there a roadmap for exposing
/rest/finance/*(or a finance-specific API surface) as a public, API-key-authenticated endpoint? The schema I care about:/rest/finance/quote/{symbol}→ real-time quote/rest/finance/timeline/{symbol}/entries→ daily narratives
most valuable/rest/finance/analyst-ratings/{symbol}→ ratings + history/rest/finance/earnings/{symbol}→ earnings with highlights
- Is this intentionally product-siloed (Finance is a consumer product, not on the API roadmap)? If so, happy to know so I stop asking.
- Any existing workaround I’m missing? Is there a way to authorize
sonar-deep-researchor Pro Search to actually fetch these specific pages via cookies/auth?
For context: I’ve already integrated sonar-pro with search_mode: "sec" for SEC filings (works beautifully, thanks!) and the general /search API for broader sources. The Finance page data is the last and hardest piece — and it’s the most differentiated, since nowhere else produces daily narratives with
Perplexity’s synthesis quality.
Would genuinely pay a premium for a /finance/* API tier. Thanks for any guidance!