Jupiterp API Docs (pre-release)

Introduction

Welcome to the Jupiterp API, a free and open-source API to get detailed course data for the University of Maryland. Currently, the API is in pre-release phase and is unstable; expect breaking changes, but the information in these docs should be correct and up-to-date.

For any questions or bugs, please contact admin@jupiterp.com.

Feel free to view or contribute to the project on GitHub.

Versions

Everything is served under /v1. Use it for new work.

/v0 still works and is not going away. Every read endpoint documented here answers on both prefixes, from the same handlers, returning the same bytes — /v0/courses and /v1/courses are the same endpoint. Existing clients, including @jupiterp/jupiterp 1.x, need no change and there is no migration deadline.

The two prefixes are held together by a parity check that compares every read endpoint across both, so they cannot quietly drift apart.

Endpoints

/v1/

(back to endpoints)

This is the base endpoint for the Jupiterp API. It will simply return a HTTP StatusOK with some text to indicate that the Jupiterp API is online.

/v1/courses

(back to endpoints)

Gets a list of courses that match the given query parameters. This endpoint does not return section information; for section info, use the sections endpoint listed below.

Query parameters

Output

Examples

Getting multiple specific courses

Request: GET http://api.jupiterp.com/v1/courses?courseCodes=CMSC131,MATH141

Response:

[
  {
    "course_code": "CMSC131",
    "name": "Object-Oriented Programming I",
    "min_credits": 4,
    "max_credits": null,
    "gen_eds": null,
    "conditions": [
      "Corequisite: MATH140. ",
      "Credit only granted for: CMSC131, CMSC133 or CMSC141."
    ],
    "description": "Introduction to programming and computer science.
    Emphasizes understanding and implementation of applications using
    object-oriented techniques. Develops skills such as program design and
    testing as well as implementation of programs using a graphical IDE.
    Programming done in Java."
  },
  {
    "course_code": "MATH141",
    "name": "Calculus II",
    "min_credits": 4,
    "max_credits": null,
    "gen_eds": null,
    "conditions": [
      "Prerequisite: Minimum grade of C- in MATH140."
    ],
    "description": "Continuation of MATH140, including techniques of
    integration, improper integrals, applications of integration (such as
    volumes, work, arc length, moments), inverse functions, exponential and
    logarithmic functions, sequences and series."
  }
]

Getting courses that satisfy Gen-Ed requirements

Request: GET http://api.jupiterp.com/v1/courses?genEds=DVUP,DSSP&limit=2&sortBy=courseCode.asc

Response:

[
  {
    "course_code": "AAST351",
    "name": "Asian Americans and Media",
    "min_credits": 3,
    "max_credits": null,
    "gen_eds": [
      "DSSP",
      "DVUP"
    ],
    "conditions": [
      "Credit only granted for: AAST351, AAST398M or AAST398N. ",
      "Formerly: AAST398M, AAST398N."
    ],
    "description": "From yellow peril invaders to model minority allies, Asian 
    Americans have crafted their own dynamic cultural expressions in a number 
    of media from film, television, and music to fashion, sports, and food that
    reveal and contest the contradictions of the U.S. nation-state. Asian
    American culture also uniquely sits at the nexus of immigration flows and
    digital technologies, providing a transnational lens to view the US place
    in the world. This advanced course, then, will introduce students to the
    study and practice of Asian American culture as multiple , hybrid, and
    heterogeneous. It will do so through three sections: section one will
    introduce students to classical, cultural, and media concepts as well as
    relevant keywords outlined by Asian American Studies scholars; section two
    will review the work of Asian American cultural theorists; section three
    will focus on analyses of particular Asian American cultural productions.
    In doing so, students will gain an understanding of the shifting and
    interlocking tensions among the local, the national, and the global that
    form the cultural geographies of Asian America."
  },
  {
    "course_code": "AMST320",
    "name": "(Dis)ability in American Film",
    "min_credits": 3,
    "max_credits": null,
    "gen_eds": [
      "DSHU",
      "DSSP",
      "DVUP"
    ],
    "conditions": [
      "Credit only granted for: AMST320 or AMST328X. ",
      "Formerly: AMST328X."
    ],
    "description": "Explores the connection between film and disability
    through an analysis of independent and mainstream American films in various
    film genres. Specifically, we will consider how these film representations
    reflect and/or challenge the shifting social perspectives of disability
    over the 20th and 21st centuries.  Beginning with the presentation of
    disability as theatrical spectacle in the traveling sideshow and early
    cinema, we will work our way through film history to develop an
    understanding of our society's complicated relationship with disability."
  }
]

/v1/courses/minified

(back to endpoints)

Gets a minified list of courses that satisfy the given parameters. Takes the same parameters as the /v1/courses endpoint, but returns only the course code and title.

Query parameters

Same as the parameters for /v1/courses; see here.

Output

Examples

Getting courses with a specific prefix

Request: GET http://api.jupiterp.com/v1/courses/minified?prefix=ASTR4&sortBy=name.asc

Response:

[
  {
    "course_code": "ASTR422",
    "name": "Cosmology"
  },
  {
    "course_code": "ASTR421",
    "name": "Galaxies"
  },
  {
    "course_code": "ASTR498",
    "name": "Special Problems in Astronomy"
  }
]

/v1/courses/withSections

(back to endpoints)

Gets a list of full courses data and associated sections data. Each returned course also contains a (potentially-empty) list of sections for that course.

Query parameters

Output

Examples

Getting a course with sections data

Request: GET http://api.jupiterp.com/v1/courses/withSections?courseCodes=CMSC433

Response:

[
  {
    "course_code": "CMSC433",
    "name": "Programming Language Technologies and Paradigms",
    "min_credits": 3,
    "max_credits": null,
    "gen_eds": null,
    "conditions": [
      "Prerequisite: Minimum grade of C- in CMSC330; or must be in the
      (Computer Science (Doctoral), Computer Science (Master's)) program. ",
      "Restriction: Permission of CMNS-Computer Science department."
    ],
    "description": "Programming language technologies (e.g., object-oriented
    programming), their implementations and use in software design and
    implementation.",
    "sections": [
      {
        "holdfile": 0,
        "meetings": [
          "TuTh-11:00am-12:15pm-CSI-1115"
        ],
        "sec_code": "0101",
        "waitlist": 3,
        "open_seats": 0,
        "course_code": "CMSC433",
        "instructors": [
          "Anwar Mamat"
        ],
        "total_seats": 140
      },
      {
        "holdfile": null,
        "meetings": [
          "TuTh-3:30pm-4:45pm-IRB-0318"
        ],
        "sec_code": "0201",
        "waitlist": 0,
        "open_seats": 7,
        "course_code": "CMSC433",
        "instructors": [
          "Anwar Mamat"
        ],
        "total_seats": 50
      }
    ]
  }
]

/v1/sections

(back to endpoints)

Get sections for specific courses, or for all courses that match a course code prefix. Note that some courses don't have any sections; for example, most independent research courses, like ASTR498, will not return any sections.

Query parameters

Output

Examples

Getting all sections for a course

Request: GET http://api.jupiterp.com/v1/sections?courseCodes=CMSC433

Response:

[
  {
    "course_code": "CMSC433",
    "sec_code": "0101",
    "instructors": [
      "Anwar Mamat"
    ],
    "meetings": [
      "TuTh-11:00am-12:15pm-CSI-1115"
    ],
    "open_seats": 0,
    "total_seats": 140,
    "waitlist": 7,
    "holdfile": 0
  },
  {
    "course_code": "CMSC433",
    "sec_code": "0201",
    "instructors": [
      "Anwar Mamat"
    ],
    "meetings": [
      "TuTh-3:30pm-4:45pm-IRB-0318"
    ],
    "open_seats": 15,
    "total_seats": 50,
    "waitlist": 0,
    "holdfile": null
  }
]

/v1/instructors

(back to endpoints)

Get a list of all instructors and their average ratings, including instructors not actively teaching any courses.

Query parameters

Output

Examples

Getting high-rated, non-5 star instructors

Request: GET http://api.jupiterp.com/v1/instructors?ratings=gt.4.5&ratings=lt.5&limit=5&sortBy=average_rating.desc,name.desc

Response:

[
  {
    "slug": "gramlich_meredith",
    "name": "Meredith Gramlich",
    "average_rating": 4.9667
  },
  {
    "slug": "cropper",
    "name": "Maureen Cropper",
    "average_rating": 4.9474
  },
  {
    "slug": "gruber_sean",
    "name": "Sean Gruber",
    "average_rating": 4.9398
  },
  {
    "slug": "o’brien",
    "name": "Terrence O’Brien",
    "average_rating": 4.9375
  },
  {
    "slug": "zomback",
    "name": "Jenna Zomback",
    "average_rating": 4.9355
  }
]

/v1/instructors/active

(back to endpoints)

Get all instructors that are currently teaching a course, as listed on Testudo.

Query Parameters

Same as /v1/instructors; see here.

Output

Same as /v1/instructors; see here.

Examples

Getting instructors currently teaching a course

Request: GET http://api.jupiterp.com/v1/instructors/active?limit=5

Response:

[
  {
    "slug": "abadi_daniel",
    "name": "Daniel Abadi",
    "average_rating": 3.122
  },
  {
    "slug": "abasi",
    "name": "Ali Abasi",
    "average_rating": null
  },
  {
    "slug": "abbasi",
    "name": "Hossein Abbasi",
    "average_rating": 3.7791
  },
  {
    "slug": "abdul-alim",
    "name": "Jamaal Abdul-Alim",
    "average_rating": 3.25
  },
  {
    "slug": "abioye",
    "name": "Victor Abioye",
    "average_rating": null
  }
]

/v1/deptList

(back to endpoints)

Get a list of 4-letter department codes.

Query parameters

None

Output

/v1/grades

(back to endpoints)

Gets grade distributions for individual course sections, as released by the University Registrar under the Maryland Public Information Act. Data covers fall and spring terms from Fall 2010 through Spring 2026; winter and summer terms were not released.

For counts aggregated across sections, terms, or instructors, use the summary endpoint below.

Query parameters

Output

Examples

Getting every section of a course in one term

Request: GET http://api.jupiterp.com/v1/grades?courseCodes=CMSC132&term=eq.202408&limit=2

Response:

[
  {
    "term": 202408,
    "course_code": "CMSC132",
    "sec_code": "0101",
    "instructor": "Herman, Larry",
    "instructor_name": "Larry Herman",
    "instructor_source": "reported",
    "total": 32,
    "a_plus": 0,
    "a": 1,
    "a_minus": 4,
    "graded": 30,
    "gpa": 2.583
  },
  {
    "term": 202408,
    "course_code": "CMSC132",
    "sec_code": "0102",
    "instructor": null,
    "instructor_name": "Larry Herman",
    "instructor_source": "lead",
    "total": 34,
    "a_plus": 2,
    "a": 7,
    "a_minus": 5,
    "graded": 34,
    "gpa": 3.118
  }
]

Note the second record: the release lists the instructor once against the lecture and leaves the discussion sections blank, so instructor is null while instructor_name carries the lecturer's name and instructor_source records that it was inferred.

/v1/grades/summary

(back to endpoints)

Gets grade distributions with the individual sections summed together. This is usually the endpoint you want: groupBy=course answers "how hard is this course", groupBy=term answers "has it changed", groupBy=instructor answers "who should I take it with", and groupBy=instructorOverall answers "how does this professor grade in general".

Note that instructorOverall and instructorTerm aggregate across every course, so they take no course filter; passing courseCodes, prefix, or number with them returns 400 rather than silently ignoring the filter.

Query parameters

Output

All groupings return the summed grade buckets (a_plus through other), total, graded, and gpa, defined exactly as on /v1/grades. In addition:

Examples

How hard is a course, over its whole history

Request: GET http://api.jupiterp.com/v1/grades/summary?courseCodes=CMSC351

Response:

[
  {
    "course_code": "CMSC351",
    "section_count": 97,
    "term_count": 32,
    "first_term": 201008,
    "last_term": 202601,
    "total": 14969,
    "graded": 13346,
    "a_plus": 450,
    "a": 1578,
    "a_minus": 1153,
    "b_plus": 1313,
    "b": 2169,
    "b_minus": 1441,
    "c_plus": 1263,
    "c": 1583,
    "c_minus": 1002,
    "d_plus": 185,
    "d": 846,
    "d_minus": 70,
    "f": 293,
    "w": 738,
    "other": 791,
    "gpa": 2.699
  }
]

Comparing instructors for a course

Request: GET http://api.jupiterp.com/v1/grades/summary?groupBy=instructor&courseCodes=CMSC330&minStudents=1000&sortBy=gpa.desc

Response:

[
  {
    "course_code": "CMSC330",
    "instructor": "Michael W. Hicks",
    "section_count": 37,
    "term_count": 7,
    "first_term": 201301,
    "last_term": 202101,
    "total": 1205,
    "graded": 1076,
    "gpa": 3.123
  },
  {
    "course_code": "CMSC330",
    "instructor": "Roger D. Eastman",
    "section_count": 39,
    "term_count": 5,
    "first_term": 201808,
    "last_term": 202108,
    "total": 1258,
    "graded": 1045,
    "gpa": 3.047
  }
]

/v1/grades/terms

(back to endpoints)

Gets every term for which grade data has been loaded, newest first. Takes no parameters. Useful for discovering coverage before querying, since the released data covers fall and spring only.

Output

Example

Request: GET http://api.jupiterp.com/v1/grades/terms

Response:

[
  {
    "term": 202601,
    "section_count": 6543,
    "course_count": 3213,
    "total": 168366,
    "graded": 161397,
    "gpa": 3.488
  },
  {
    "term": 202508,
    "section_count": 7056,
    "course_count": 3263,
    "total": 186608,
    "graded": 176931,
    "gpa": 3.503
  }
]

Jupiterp API v1 (reviews)

The endpoints below write, and they are governed differently from the read endpoints above even though both are served under /v1. Writes need things reads do not: an origin allowlist, authentication, rate limiting, and a captcha. The read endpoints stay permissive, unauthenticated, and cacheable.

That difference is worth stating plainly, because it is the one thing the shared prefix hides. Reads accept requests from any origin. Writes accept them only from an allowlist (V1_ALLOWED_ORIGINS). A browser on an unrelated domain can call GET /v1/courses and will be refused by POST /v1/reviews.

Reviews are pre-moderated. Nothing submitted here is publicly visible until a moderator approves it, and that is true whether the decision is made by a person or by the automated triage.

GET /v1/reviews

Approved reviews only, newest first. Served from a database view that cannot express an unapproved row and does not contain the submitter's identity columns at all.

The total is returned in the Content-Range header.

POST /v1/reviews

{
  "instructor_slug": "shane-walsh",
  "course_code": "CMSC132",
  "term": 202508,
  "rating": 4.5,
  "expected_grade": "A-",
  "title": "Genuinely excellent lecturer",
  "body": "…",
  "email": "student@terpmail.umd.edu",
  "captcha_token": "0.abc…"
}

rating is a decimal between 1 and 5 on a half step — 4.5 is valid, 4.3 is not. email must be a terpmail.umd.edu or umd.edu address; it is stored only as a peppered hash, is never displayed, and is never shown to the professor. course_code and term are optional, and term must be a Fall or Spring term, because the grade dataset covers only those.

Responds 202 Accepted with {"status":"verification_sent"}.

The response is identical whether or not that address has already reviewed this professor. A distinguishable "you have already reviewed this" would turn the endpoint into an oracle for "did person X review professor Y", which is the privacy property the hashing exists to provide.

Rate limited to 5 per hour per IP, 3 per day per address, and 20 per hour per professor across all submitters. The last one is what catches a coordinated run on a single professor, which the per-person limits do nothing about.

GET /v1/reviews/verify/:token

Confirms the emailed link, moves the review to pending, and returns the manage key once:

{ "status": "verified", "manage_key": "…", "message": "…" }

Idempotent: a second visit returns already_verified rather than an error, because mail clients prefetch links and people double-click.

The manage key is also emailed. It cannot be recovered — there is deliberately no way to link it back to a person.

DELETE /v1/reviews/:id

Authorization: Bearer <manage key>.

A withdrawal is a soft delete — the row remains so the one-review-per-person rule still holds, but the content is actually nulled.

There is no edit endpoint. A published review is final text: the only way to change what a review says is to withdraw it and write another. Editing after approval is a way to get innocuous text past a moderator and then replace it, and re-queueing every edit for moderation solves that at the cost of a flow where a reviewer can silently republish. Withdrawal carries no such hole, so it is the one the reviewer keeps.

PUT /v1/admin/reviews/:id

{
  "action": "approve",
  "reason": "…",
  "confidence": 0.93,
  "categories": [],
  "policy_version": "2026-08-14",
  "model": "gemini-2.0-flash-001"
}

Two callers with different keys: a human moderator with the admin key, and the automated triage with a narrowly scoped callback key that authorises this one route. Which one acted is recorded on every decision.

Idempotent — asking for the state a review is already in is a success, not a second audit entry. State-guarded — only pending and escalated reviews are decidable, and a late retry against a review a human already actioned returns 409 rather than overturning it.

Every call writes an audit row. While shadow mode is on, an automated decision is recorded with applied: false and the review is escalated to a human instead.

Errors