Clinical trial research in minutes, checked against the source.
The trial research an analyst spends days on: search ClinicalTrials.gov and PubMed, see the landscape at a glance, and compare it with your own protocol. Every number traces back to a registry record or paper, and it all runs privately on your own computer.
From search to answer in minutes
Pull the trials, count them, read the records and write it up, with every source attached.
Search
Query ClinicalTrials.gov and PubMed, separately or together, by condition, drug, sponsor, phase, status, country and dates. Trials are linked to the papers that report them.
See the landscape
Exact charts of phases, statuses and sponsors, a readout timeline of upcoming primary completions, and a breakdown of why trials stopped. Download any chart for your slides.
Ask
Summarize the landscape, compare trials side by side, or ask your own questions. Answers come from a model running on your computer and cite every trial and paper they use.
See it run
A real session, sped up: search, compare three trials, switch the internet off, attach a confidential protocol, and save the answer as a PDF.
Built for answers you can defend
Your protocol against the competition, offline
Attach a draft protocol, switch the internet off, and compare its eligibility criteria with the competing trials recruiting the same patients. The analysis runs on your computer, so the protocol never goes anywhere.
- See which of your criteria narrow the patient pool the most
- Works the same with data-room documents and client briefs
Every citation checked
TrialCarta checks every NCT ID, PMID and file name in an answer against the records you retrieved, and flags anything it can't find. Click any ID to see the record the model read.
- Numbers that the cited records don't contain are listed to double-check
- Background knowledge is labeled as such, not dressed up as a source
Trial names hidden for this page; in the app, each row links to its record.
Evidence, not predictions
TrialCarta doesn't guess how a trial will turn out. It shows how comparable trials actually played out, with the trials behind every number.
For example: completed phase 2 MASH trials took a median of 21 months from start to primary completion, and six competitor readouts have passed their registered date without an update.
Know what reads out when
A timeline of ongoing trials from start to primary completion, colored by phase. Copy it straight into a deck.
Saved searches show what changed since your last look: new trials, status changes, moved readout dates, reworded endpoints and new results.
See why trials stopped
TrialCarta groups the terminated, withdrawn and suspended trials in any search by the reason each sponsor registered: enrollment, efficacy, safety, business, funding and more.
Every chart is counted from the records, never estimated by AI, and the write-up quotes each sponsor's own words.
A one-page brief to share
Export the landscape as a one-page PDF: counts, phases, sponsors, next readouts and why trials stopped. It has no AI-written text, so it's safe to send outside the team.
Answers save as PDFs too, labeled as AI-written, with their citation check.
Example with a fictional sponsor, drug and registry record (Northwind Bio, NWB-204).
Check a protocol against its registration
Attach a trial's protocol and TrialCarta lines up its endpoints, sample size, arms and phase with the ClinicalTrials.gov registration, listing every difference with both sources cited.
- Diligence on a target's protocol from the data room, offline
- Check your own registration before it goes public
Trial names hidden for this page; in the app, each row links to its record.
Did they post their results?
FDAAA 801 expects results within a year of primary completion. TrialCarta checks every finished trial in a search: posted, in review, delay requested, or nothing at all.
For example: of 144 MASH trials whose results were due, 105 have posted them and 11 FDA-regulated trials show nothing, some for years. These are flags to review, computed from registry fields, not legal findings.
For study teams and sites
The document work that eats a CRA's or coordinator's week. Done by rules, not AI, and the files never leave your computer.
Compare protocol versions
Drop in two versions of a protocol and see every change to eligibility, dosing, visits, safety reporting and sample size, with the section and page on both sides. It compares the full text, so it also catches changes the summary of changes doesn't mention.
Example with a fictional protocol. Click to enlarge.
Check redaction before you send
Catches names, MRNs and dates of birth that are still readable: missed entirely, under a black box drawn over live text, struck through, legible in a scan, or sitting in a spreadsheet column. The report masks every value it finds.
Example with made-up test documents. Click to enlarge.
How to check that a PDF is really redactedCompare eligibility
Age, stage, labs, prior treatments and exclusions across trials, and against your own protocol.
Side-by-side comparisons
Up to six trials in one table, with the differences that matter.
Your documents, always current
Attach files, or link a folder and TrialCarta reads new and changed files before every question: PDFs, scans, Word, CSV and text.
Ready for your report
Paste into Word or email with tables intact, save as PDF, or export to CSV.
Projects that remember
Each project keeps its searches, documents and chats, so you pick up where you left off.
Built around your data
Your own data source, a report in your format, a check your team runs every week: tell us what you need and we'll build it for your team, running on your computers like the rest of TrialCarta.
For teams: one Mac, everyone's browser.
Install TrialCarta on one Mac and turn on Share with my team. Colleagues on Windows, Mac or Linux sign in from Chrome, Edge, Safari or Firefox, with nothing to install on their computers.
- Accounts by invitation. An administrator invites each person and they set their own password. Chats are private to whoever asked.
- Shared projects. One per study or client, with its documents and saved searches, open to the people you choose.
- Internet searches under your control. Turn ClinicalTrials.gov and PubMed searches off for the whole team, or just for a confidential project.
- Everything stays on that Mac. Documents, questions and answers never leave it, and the AI model runs there for everyone.
- Priced per person. A yearly license for the number of people on your team.
The Team screen in TrialCarta, with example addresses.
For biotech finance: who needs capital before the readout.
TrialCarta Pro reads ClinicalTrials.gov, SEC EDGAR and PubMed together and ranks drug developers with a readout coming, listed and private, so royalty, credit and BD teams see the next deals early.
- Financing Screen. A ranked list of drug developers with readout timing, cash and runway, and whether the cash lasts to the readout. Analytics, CSV and a PDF brochure.
- Private companies too. What SEC filings say about how privately held sponsors are funded.
- New companies. Sponsors whose first trials were just posted, how they're funded and what's published on the drug. Keep the ones you like on your radar.
- Alerts. Weekly re-checks that flag new raises, readouts that moved, trials stopped or completed, results posted and royalty deals signed.
- Deal terms from filings. Licensing, royalty and financing agreements on EDGAR, with the upfront, milestone and royalty terms quoted from the filing.
- Briefs on your Mac. Ask for a brief on any company: every figure cites its filing or trial record, and your questions never leave the building.
Tested on history: run as of mid-2021 and early 2023, the ranking put two to three times more of the companies that went on to sell a royalty in its top ten than a random pick would. Public data; a screening aid, not investment advice.
Financing Screen analytics from a live run on public data, company names left out.
Confidential work stays on your computer.
TrialCarta runs on your own computer, not a website. The AI runs there too, so a protocol, a data-room document or a client brief is never uploaded anywhere.
- One switch blocks the internet. Turn Internet Access off and the app and its helpers can't reach the network at all, enforced by a macOS sandbox, not a promise.
- Open models, running locally. No cloud AI service sees your questions or your files. Internet is used only to search public registries and to download models.
- No usage limits or per-query costs. Nothing is metered, and there's no cloud AI vendor to approve.
- No cloud account, no tracking. Your projects and chats are files on your computer, and deleting a project deletes its data. Team accounts live only on your own Mac.
- Runs on Apple silicon Macs today. Works with 16 GB of memory; with 32 GB it uses a larger model for the best answers.
- For a team, one Mac serves everyone. Colleagues use it from their browser on Windows, Mac or Linux, and documents never leave your network. How it works for teams.
Who it's for
CRAs, coordinators and sites
See what changed between protocol versions, turn the schedule of assessments into a visit checklist, check redaction before documents go out, and know which trials nearby recruit the same patients.
Clinical operations and feasibility
Size up competing trials, compare eligibility against your draft protocol, and learn from trials in your indication that stopped for slow enrollment.
Competitive intelligence
Track who's ahead in each indication, what reads out next, and what changed this week, with every claim traceable to a registry record or paper.
BD, investors and diligence
Assess an asset's landscape and risks quickly, find companies that may need capital before their readout, and work with confidential deal documents without sending them to a cloud service. TrialCarta Pro.
TrialCarta supports research. It is not medical advice.
bRAGs: chat with your own documents, privately
Ask questions about PDFs (including scanned ones), Word files, images or a whole folder, all on your computer. Point it at a project folder and it keeps up with your changes before every question.
See TrialCarta on your own indication.
We're based in San Diego. Tell us the indication you care about and we'll show you TrialCarta working on it.
Prefer email? Write to our team.


