Delv
Researchby Fiscal.ai4.3

Fiscal.ai (FinChat)

Conversational investment research agent for retail and institutional investors covering 100,000+ global companies with earnings transcripts and filings.

B
Safety & Trust

Delv Safety Grade: B

Score 72/100 · assessed 2026-04-18

Maintainer65
Permissions85
Supply chain60
Transparency55
Incidents100

Fiscal.ai (FinChat) is a web-based conversational research agent from a specialist fintech startup, not a major vendor. The service operates entirely in-browser with read-only access to financial databases (SEC filings, earnings transcripts), which limits blast radius compared to filesystem or shell tools. No repository is public, so you cannot audit the underlying code or dependencies. The freemium model and web-only delivery reduce supply-chain risk versus installable packages, but transparency suffers from closed-source architecture and minimal public documentation of data sources or AI model provenance. No known security incidents. Permissions are appropriately scoped to network outbound (API calls for financial data) and potential external LLM usage. Suitable for research workflows where you trust the vendor's data handling, but institutional users should verify compliance with internal data policies before feeding proprietary queries.

Green flags

  • Read-only financial data access minimises write-based risks
  • Web-only delivery avoids local filesystem or shell exposure
  • Covers 100,000+ companies with cited filings, reducing hallucination risk
  • No known security incidents or credential leaks to date
  • Scoped to investment research domain, not general-purpose automation

Red flags

  • No public repository or source code available for audit
  • Closed-source architecture obscures data handling and model behaviour
  • Unclear which LLM provider processes user queries and company data
  • Freemium model may incentivise data monetisation beyond stated use
  • Limited transparency on data retention and query logging practices

Permissions requested

Outbound networkExternal LLM call
Assessed by Delv Editorial using public metadata. Grades are advisory and update as the ecosystem changes. They do not replace your own review of permissions and code before granting an agent access to sensitive systems.

Pricing

FREEMIUMFree tier, paid from ~$20/mo

Platforms

web

Review

Fiscal.ai positions itself as a conversational research agent for investors, and the autonomy claim hinges on its ability to surface relevant earnings transcripts, SEC filings, and financial metrics without you manually hunting through databases. You ask a question like "How did Nvidia's data centre revenue compare to AMD's last quarter?" and it pulls the numbers, cites the filings, and presents a comparative table. The agent shines when you need quick synthesis across multiple companies or quarters. I've used it to track margin trends for a basket of SaaS companies, and it saved me from opening ten different investor relations sites. The conversational layer is genuinely useful here, it remembers context across follow-ups, so you can drill into gross margin, then ask about R&D spend, without restating the ticker list. Where it earns its keep over a standard AI tool is the structured data layer. It's not hallucinating revenue figures because it's pulling from a curated dataset of 100,000+ companies. That's the autonomy: it knows where to look and what to cite. Compared to Bloomberg Terminal or FactSet, it's vastly cheaper and more approachable for retail investors or small fund managers. Compared to ChatGPT with web search, it's more reliable and doesn't require you to verify every number against EDGAR. Failure modes: it's still a chatbot, so vague questions get vague answers. Ask "Is Tesla a good buy?" and you'll get a summary of recent performance, not actionable insight. It won't build you a DCF model or run scenario analysis autonomously. The free tier is limited, you hit usage caps quickly if you're doing serious research. The interface is web-only, no API or desktop app, which feels limiting for power users who want to pipe data into spreadsheets or custom dashboards. It also skews heavily toward US equities; international coverage exists but feels thinner. One concrete workflow: I use it to prep for earnings calls. I ask for the last four quarters of revenue and EPS, then request management commentary on a specific segment. It pulls the transcript excerpts, and I've got a briefing doc in five minutes. That's the value proposition, speed and citation quality for routine research tasks. If you're a retail investor doing your own DD or a junior analyst who needs to ramp up fast on a sector, this is worth the $20/month. If you need deep quant analysis or custom screening, look elsewhere.
Verdict

Pay for this if you're a retail investor or small fund analyst who needs fast, cited financial data without Bloomberg's price tag. Skip it if you need custom models, API access, or deep international coverage.

Good at

  • Pulls from curated dataset of 100,000+ companies, minimises hallucination risk on financials
  • Conversational memory lets you drill down across follow-ups without restating context
  • Cites earnings transcripts and filings directly, saves manual EDGAR trawling
  • Freemium tier lets you test before committing to $20/month
  • Faster than Bloomberg for routine research, vastly cheaper for retail users

Watch out

  • Web-only interface, no API or desktop app for power users
  • Free tier usage caps hit quickly if you're doing serious research
  • Won't build financial models or run scenario analysis autonomously
  • International equity coverage thinner than US stocks
  • Vague questions yield vague answers, requires specific prompts to shine

Use cases

  • stock research
  • earnings analysis
  • portfolio tracking