Multi-agent AI for academic applications

PhD applications,
handled end-to-end.

ACAIS is a multi-agent system that discovers PhD programs and scholarships across the web every 6 hours, scores each one against your research profile, writes tailored statements and cold emails, then sends them — under whatever level of control you choose. Upload your CV once. The agents run the rest.

Get startedHow it works
What it does
01
Finds programs for you
Searches academic platforms every 6 hours for new PhD programs and scholarships — automatically, without you having to check anything.
02
Scores each one against your background
Compares every program to your education, skills, and research interests and gives it a score out of 100. The higher the score, the better the fit.
03
Writes your application materials
Writes your statement of purpose and the email to the supervisor — tailored to each specific program. You can read and edit before anything is sent.
04
Sends when you say so
You decide whether to review everything first, approve with one click, or let it send on its own. Either way, you always see a log of what was sent and when.
7
Specialized agents
6h
Discovery cycle
5+
Academic platforms scraped
0–100
Match score per program
Search AgentExtraction AgentProfile AgentMatching AgentSupervisor AnalysisWriting AgentExecution AgentGPT-4oFAISS Vector SearchAutomated ApplicationsStatement of PurposeCold Email WritingApplication Tracking
Search AgentExtraction AgentProfile AgentMatching AgentSupervisor AnalysisWriting AgentExecution AgentGPT-4oFAISS Vector SearchAutomated ApplicationsStatement of PurposeCold Email WritingApplication Tracking
01. How it works
07 AI modules

ACAIS is made up of seven AI modules, each responsible for one part of the application process. They run in sequence — no action required from you unless you want it.

01→
Finds programs
while you sleep.
Every 6 hours, the AI searches FindAPhD, DAAD, NatureCareers, and other academic platforms for new PhDs, scholarships, and research roles. You don't have to check anything manually — it runs on its own.
FindAPhDDAADNatureCareersRuns every 6h
02→
Reads every
listing for you.
When a new program is found, the AI visits the page and reads the full details — requirements, funding, supervisor info, deadlines — and saves it in a structured format so it can be searched and compared.
Automated browsingAI readingStructured data
03→
Understands
your background.
Upload your CV as a PDF or Word file. The AI reads it and builds a complete picture of your education, skills, research interests, and publications — so it knows what kind of programs you would be a good fit for.
PDF uploadWord uploadAutomatic parsing
04→
Shows you what
actually fits.
The AI compares your profile against every discovered program and gives each one a score from 0 to 100. It considers your research background, skills, funding needs, and how well your interests match the supervisor.
0–100 match scoreResearch alignmentSkill fit
05→
Researches who
you're writing to.
Before writing anything, the AI looks up the supervisor — their recent papers, research focus, and academic standing — and uses this to make sure your application speaks directly to their work.
Google ScholarRecent publicationsResearch themes
06→
Writes your
application.
The AI writes a tailored statement of purpose (600–800 words) and a cold email to the supervisor, referencing one of their actual published papers by name. It also reviews your CV and tells you what to improve.
Statement of purposeCold emailCV review
07→
Sends it
on your terms.
You choose how much control you want. Review everything before it goes out, approve with one click, or let it run fully on its own with a daily safety limit. Every action is logged so you always know what was sent.
Manual reviewOne-click approve10 emails/day limit
02. The pipeline
04 stages
01.
Upload your CV.
The rest starts.
The moment you upload your CV, the AI reads it and builds your profile. At the same time, discovery runs in the background — searching academic platforms every 6 hours and adding new programs to the index automatically.
Automatic discoveryCV parsingRuns every 6h
02.
It learns
who you are.
Your education, skills, research interests, and publications are extracted and understood. The AI builds a complete picture of your background so it knows exactly what kind of PhD or scholarship you would be a strong candidate for.
SkillsResearch interestsPublicationsCompleteness score
03.
You see what
fits you best.
Every program in the index is scored against your profile from 0 to 100. The score reflects how well your research background, skills, and interests line up — including how well you match the specific supervisor.
0–100 scoreResearch alignmentSupervisor fit
04.
Application written
and sent.
The AI writes a tailored statement of purpose and a cold email — with a reference to one of the supervisor's actual papers. You review it, approve it, or let it send automatically. Your choice, every time.
Statement of purposeCold emailReview or auto-send
03. The idea
Most PhD applicants spend months writing the same email to a hundred supervisors.
ACAIS does it properly — one profile, seven agents, every program scored and written for.
04. Automation modes
03 modes
01
You review
everything.
Agents prepare all materials and hold them for your review. Nothing is sent without explicit confirmation.
All drafts held for reviewNo automated sendsFull step visibility
02
Recommended
One click
to send.
Agents draft the SOP and cold email. A single approval button is all that stands between the draft and the supervisor.
Auto-generate materialsOne-click approveRecommended default
03
It runs
itself.
The full pipeline runs autonomously. A 10 emails/day hard cap and an audit log keep it safe and auditable.
Fully autonomous10 emails/day capImmutable audit trail
05. Stack
09 layers
APIFastAPIPython async, Uvicorn, auto-generated OpenAPI docs at /api/docs
DatabasePostgreSQL 15SQLAlchemy 2 async, Alembic migrations, UUID primary keys
Task queueCelery 5Redis broker, Beat scheduler, Flower task monitor at :5555
Vector DBFAISSIn-process index using text-embedding-3-small, Pinecone-ready
AIGPT-4oExtraction, profile parsing, document generation, scoring
ScrapingPlaywrightHeadless Chromium, BeautifulSoup, httpx, rate-limited
AuthJWTpython-jose, bcrypt, refresh token rotation
FrontendNext.js 14App Router, React Query, Tailwind CSS, typed Axios client
ContainersDocker ComposeOne command starts all services: database, Redis, API, workers, frontend
06. Get started
Upload your CV.
The agents take it from there.

Create an account, upload your CV, and ACAIS builds your research profile, discovers programs that match it, and prepares tailored application materials — all within minutes of your first upload. No manual searching. No writing from scratch.

Open ACAISAPI docs