Earnings season is not a reading problem. It is a triage problem.

Roughly 10,000 companies report in a single quarter — the last full 90-day window put just under that number of calls into the archive I maintain, and the busiest single day of the spring carried 877. Even a focused analyst with 60 names on the sheet is looking at 60 calls inside three weeks, each 8,000 to 12,000 words, most stacked on the same four Tuesday and Thursday mornings.

So the real job is not reading. It is deciding what to read closely. Which three calls deserve a full pass with a highlighter, which twenty need a targeted search for one line about pricing or backlog, and which thirty-seven you clear with a summary and a glance at the numbers. Get the triage right and you have time for genuine work on the calls that matter. Get it wrong and you spend the season skimming everything and understanding nothing — the failure mode I wrote about here.

Tools help in different ways. Some tell you what happened to the numbers, some tell you what management said, some get you the call live, and some sit on top and do the reading with you. Below are nine a working analyst would actually consider, ordered by how much of the season they carry. I built the first one, so read that ranking with the appropriate amount of salt — but I have been honest about what the others do better than we do.

1. earningscalls.dev

Best for: searching what was said, across companies and across quarters, without a seat licence.

This is the transcript layer. 253,000+ earnings call transcripts covering 12,799 companies across 175 exchanges in 70+ countries, from 2020 to today, split into 11.99 million speaker-tagged segments. Every transcript is free to read on the site, with no account and no article limit. For a season specifically, four things carry the weight: an upcoming-earnings calendar, speaker-role filtering that separates scripted prepared remarks from the analyst Q&A where the useful friction lives, full-text search across the entire corpus so you can compare what a company said this quarter against its own prior quarters and against peers reporting the same week, and a native MCP server so Claude can do the reading alongside you. Paid tiers cover API and volume access — Pro at $24.99, Ultra at $39.99, Enterprise at $299 per month — but the research workflow above costs nothing.

Limitation: no live audio during the call, so you cannot trade the call itself on our data. No estimates, no consensus, no price charts — we do not have the numbers side at all, and you will need something else on this list for that. History starts in 2020, so a 2008-vintage cycle comparison is out of reach.

2. Bloomberg Terminal or FactSet

Best for: everything at once, when somebody else is paying.

These are the institutional standard for a reason. Consensus estimates with revision history, ownership, credit, supply-chain relationships, the news wire, chat with the street, and transcripts too. When a print crosses, the terminal tells you the beat or miss against the number the street actually carried — not the number a retail site scraped — and that distinction decides whether a stock is up or down eight percent. Nothing here replaces that. If you have a terminal, it is the backbone of your season and the rest of this list slots in around it.

Limitation: neither publishes pricing. It is enterprise pricing, negotiated per seat on multi-year contracts, and the figures you see quoted online are third-party estimates rather than anything the vendors confirm. Beyond cost: the data is seat-locked. Getting transcript text out of a terminal and into your own scripts or an LLM is deliberately hard, which is precisely the gap the cheaper tools below fill.

3. AlphaSense

Best for: searching across every document type at once, not just transcripts.

AlphaSense beats us comprehensively at breadth. It indexes filings, broker research, trade press, company presentations and expert-call transcripts in one search surface with summarisation on top, and for a team that needs "everything anyone has written about this supply chain" it is the correct tool. Its coverage of paid broker research is something no free product will ever match, because that content is licensed.

Limitation: enterprise pricing, annual contracts, per seat, with no public price list and no realistic path for an individual investor. Reported seat costs run into five figures per year and rise sharply for the expert-transcript tiers. Licensing also restricts what you can export, so it is a place you go to read, not a source you pipe into your own models.

4. Quartr

Best for: live events and genuinely global coverage, with a free mobile app.

Quartr is the best-known consumer-facing way to actually listen to a call, with live audio, slides, filings and transcripts in one app across 65+ markets. The mobile app is free and very good — for following a call in real time on a phone it is hard to beat. There is also an API and a professional desktop product aimed at institutions and AI companies that want structured IR data as infrastructure.

Limitation: the app is built for consuming events, not querying a corpus, so cross-company research is awkward. The API, exports and enterprise controls are commercial contract products with no public pricing, which makes "what will this cost me" a sales conversation rather than a checkout page.

5. Aiera

Best for: working the call live, with transcription inside seconds.

If your edge depends on reacting during the call rather than after it, Aiera is the specialist. It runs real-time transcription with sub-second delay on a finance-tuned speech model, covers tens of thousands of events a year beyond earnings — conferences, investor days, shareholder meetings — and turns around a first-draft transcript within minutes of the call ending, human-reviewed versions following. For a desk that needs the words while the stock is still moving, that speed is the entire product.

Limitation: an institutional product with pricing you have to ask for. It is also overkill if you are not trading the event itself — if you read transcripts the next morning like most investors do, you are paying for latency you will never use.

6. Koyfin

Best for: charts, fundamentals and estimates at a price an individual can actually pay.

Koyfin is the closest thing to a terminal-style numbers workspace for people without a terminal budget. Dashboards, global screening, ten years of financials and forward estimates on the paid tiers, sector and macro views, in a layout built for someone working a watchlist rather than a single ticker. Pricing is public and modest: a free tier, Plus at $39 per month and Premium at $79, with advisor tiers above that.

Limitation: it is not a transcript tool. Koyfin tells you what the numbers did and what the street expects next; it will not tell you the phrase the CFO used three times in the Q&A or how it differs from last quarter. Pair it with something here that reads words.

7. Seeking Alpha

Best for: a transcript plus a second opinion, in one place, for a retail price.

Seeking Alpha is the default for many individual investors because it bundles transcripts, quant grades and a large volume of written commentary behind one consumer subscription — Premium is $299 a year at list, frequently discounted. For a name you do not follow closely, reading the transcript and then three contributor takes on the same quarter is a fast way to find the bear case you had not considered.

Limitation: transcripts sit behind the paywall, coverage skews heavily US, and there is no bulk or programmatic access — you read one article at a time in a browser. Contributor quality is variable and the incentives reward volume and strong opinions, so treat the commentary as a prompt for your own work, not a conclusion.

8. Finviz or TradingView

Best for: deciding which names deserve attention, and seeing what price already knows.

Screening is the other half of triage. Finviz is faster for scanning a whole market on fundamentals down to a shortlist; TradingView is better once you have that shortlist and want the chart, the reaction to the last four prints and where the stock sits in its range. Both have usable free tiers and paid plans in the tens of dollars per month. In a season they answer one question: what moved, and is the move already priced.

Limitation: neither has a narrative layer. A screener tells you a stock gapped nine percent on a beat; it cannot tell you the beat came from a one-time licence payment the CFO mentioned once in the Q&A. Price context without transcript context is how people buy the wrong beat.

9. Claude or ChatGPT

Best for: the analysis layer on top of whatever data you feed it.

An LLM is the multiplier on every other tool here. Give it four quarters of executive segments and it will tell you which topics were dropped and which hedge phrases appeared. Give it analyst Q&A from five peers in the same week and it will surface the question the whole sector is being asked. Both vendors sell consumer tiers around $20 a month, the cheapest capability increase available to a researcher right now.

Limitation: an LLM is only as good as the data you put in front of it. Asked about a call from memory, it will produce a plausible, confidently wrong paraphrase and occasionally invent a quote. It needs a retrieval path — an MCP server or an API handing it the actual text — and without one it is a liability during earnings season, not a tool.

Build your stack

Three setups that actually work, at three budgets.

Retail investor, near zero cost. Transcripts, calendar and cross-quarter search on earningscalls.dev free. Koyfin free tier for charts and fundamentals. Finviz or TradingView free for screening and price context. Claude or ChatGPT on a consumer plan as the analysis layer, connected via MCP so it reads real transcripts instead of guessing. Total: zero to about $25 a month. The honest gap is consensus estimates — you work off whatever numbers your broker shows you, and those are not always the ones the market traded on.

Small fund or independent analyst, roughly $100–150 a month. Same transcript core, upgraded to Pro or Ultra for API access so you can script the boring parts. Koyfin Plus or Premium for the numbers. Seeking Alpha Premium or a broker research feed for second opinions. An LLM subscription. This covers maybe 85 percent of a terminal for under two percent of the cost. The missing 15 percent is real: consensus revision history, ownership data, and the ability to ping a sell-side analyst in chat.

Institutional, terminal already paid for. Keep the terminal for estimates, ownership, market data and the wire — nothing here replaces it. Add AlphaSense if the firm licences it. Then slot us in underneath as the searchable transcript layer, via the earnings call API or MCP. Enterprise at $299 a month is a rounding error next to one terminal seat, and it buys the thing the terminal deliberately does not: transcript text in bulk that your own models and notebooks can query, across 70+ countries, without a per-seat licence on every analyst who wants to run a screen.

The 15-minute pre-call routine

The stack matters less than the routine you run before every call. The short version: pull the last eight calls for baseline, read executive-only segments from the last four to spot what management dropped or started emphasising, read analyst-only segments from the same four to find the question that keeps coming back, search one key theme across the peers reporting that week, and write a one-page brief of five things to listen for — each with the verbatim quote you are testing against.

Fifteen minutes, and you walk into the call with a hypothesis instead of a blank page. The full version, with the exact prompts, is here: prepare any earnings call in 15 minutes with Claude and MCP.

Every tool here is good at something. None of them decides what matters in your coverage — that is still your job. What they buy you is the time to do it.

Related comparisons

Start with the transcripts. Every call, every quarter, free to read at earningscalls.dev.