Luni
Most “AI transfer advice” fails for one reason:
It’s not connected to the systems that define reality.
A chatbot can explain what ASSIST is. It cannot guarantee your plan matches your college, your target campus, your major, and your academic year — unless it can use tools to query the underlying data.
That’s the difference between a chatbot and a tool-using AI agent.
What a tool-using agent actually does
An agentic workflow looks like this:
- query the rules (ASSIST articulation, GE lists, course catalogs)
- convert rules into constraints (sequencing, eligibility gates)
- propose a term plan that is enrollable
- re-check and propose safe swaps when reality changes
Data snapshot (why tool access matters)
These are real counts returned by Luni’s public APIs (fetched on 2025-12-25). This is the “surface area” a chatbot must navigate — and why tool access is non-negotiable.
| System | What the API returns | Example (real snapshot) |
|---|---|---|
| ASSIST articulation | major-by-major agreement objects | De Anza → UC Irvine (2025–2026) exposes 90 majors via /api/v1/agreement/majors |
| Transferable course lists | course records with validity windows | De Anza UCTCA (2024–2025): 897 records; CSUTC (2024–2025): 1,560 records |
| Volunteer opportunities | time-windowed opportunity feed | California next 30 days: 1,367 events total via /api/v1/volunteer/events |
| UC admissions data | campus & major admit stats | UC major tables (2013–2023) served via /api/v1/data/*_major_table.json |
If your AI can’t query these, it’s guessing.
What this means for you (outcomes, not hype)
When an AI agent can use real tools, you get outcomes that matter:
- fewer wasted units (because the plan matches official rules)
- fewer timeline surprises (because constraints are checked early)
- higher-quality decisions (because choices are grounded in data)
- less time spent in “random tab hell”
Try the tool layer (free)
- Major rules (ASSIST): /transfer-assist
- Course lists (IGETC/UCTEL/CSUGE…): /transferable-courses
- UC admit data explorer: /uc-transfer-data
- Volunteer opportunities: /background-enhancement/volunteer-opportunities/ca
The paid move: the app is where the agent executes
Free tools prove the tool layer exists. The paid product is the execution view.
In the paid Luni app, the agent connects tools into one workflow: plan → monitor → adjust → ship your application.
-
Agent-built term plans with citations back to official rules
-
Risk flags + safe swaps when reality changes (sections fill, offerings shift, you change targets)
-
Ops-grade analytics + admit probability (est.) to focus on high-leverage moves
-
Opportunity Radar + Essay Workspace + milestones so story and execution stay synchronized
-
See pricing & join the waitlist: /pricing
-
Explore app features: /features
FAQ
Is this “agentic AI” just a buzzword?
Not here. It means the AI can call tools that return verifiable outputs (rules, lists, stats) instead of generating plausible-sounding text.
Will the plan always be correct?
You still verify final details — but tool-based planning dramatically reduces the biggest failure mode: planning against the wrong constraints.