A high-performance Python implementation of TypeIDs — type-safe, sortable identifiers based on UUIDv7.

TypeIDs are designed for modern systems where identifiers should be:

• globally unique
• sortable by creation time
• safe to expose externally
• easy to reason about in logs, APIs, and databases

This library provides a Python package with Rust acceleration.

• ✅ UUIDv7-based, time-sortable identifiers
• ✅ Schema-based ID explanations (JSON / YAML)
• ✅ Fast generation & parsing (Rust-accelerated)
• ✅ Multiple integrations (Pydantic, FastAPI, ...)
• ✅ Type-safe prefixes (user_, order_, ...)
• ✅ Human-readable and URL-safe
• ✅ CLI tools (new, encode, decode, explain)
• ✅ Fully offline, no external services

TypeID is optimized for real-world performance, not just correctness.

• 🚀 ~5× faster generation
• ⚡ ~1.6× faster parsing
• 🔁 ~2.5× faster end-to-end workflows

Benchmarks are:

• reproducible
• committed as raw JSON
• runnable locally via bench/

See Docs: Performance for details.

$ pip install typeid-python

Included:

• Rust base32 encode/decode
• uuid-utils for fast UUIDv7 generation

$ pip install typeid-python[yaml]     # YAML schema support
$ pip install typeid-python[cli]      # CLI tools

Extras are strictly optional.

from typeid import TypeID

tid = TypeID(prefix="user")

assert tid.prefix == "user"
assert isinstance(tid.suffix, str)
assert str(tid).startswith("user_")

from typeid import TypeID

tid = TypeID.from_string("user_01h45ytscbebyvny4gc8cr8ma2")
assert tid.prefix == "user"

from typeid import TypeID
from uuid_utils import uuid7

u = uuid7()
tid = TypeID.from_uuid(prefix="user", suffix=u)

assert tid.uuid.version == 7

from typing import Literal
from typeid import TypeID, typeid_factory

UserID = TypeID[Literal["user"]]
gen_user_id = typeid_factory("user")

user_id = gen_user_id()

$ pip install typeid-python[cli]

Generate:

$ typeid new -p user
user_01h2xcejqtf2nbrexx3vqjhp41

Decode:

$ typeid decode user_01h2xcejqtf2nbrexx3vqjhp41
uuid: 0188bac7-4afa-78aa-bc3b-bd1eef28d881

Encode:

$ typeid encode 0188bac7-4afa-78aa-bc3b-bd1eef28d881 --prefix user

TypeID is framework-agnostic by design. Integrations are provided as optional adapters, installed explicitly and kept separate from the core.

• Pydantic (v2) Native field type with validation and JSON Schema support.

  from typing import Literal
  from pydantic import BaseModel
  from typeid.integrations.pydantic import TypeIDField
  
  class User(BaseModel):
      id: TypeIDField[Literal["user"]]

• FastAPI (Coming Soon 🚧)

• SQLAlchemy (Coming Soon 🚧)

All integrations are opt-in via extras and never affect the core package.

$ typeid explain user_01h45ytscbebyvny4gc8cr8ma2

Outputs:

parsed:
  prefix: user
  uuid: 01890bf0-846f-7762-8605-5a3abb40e0e5
  created_at: 2025-03-12T10:41:23Z
  sortable: true

Works without schema, fully offline.

Define meaning for prefixes using JSON or YAML.

Example (typeid.schema.json):

{
  "schema_version": 1,
  "types": {
    "user": {
      "name": "User",
      "owner_team": "identity-platform",
      "pii": true
    }
  }
}

Then:

$ typeid explain user_01h45ytscbebyvny4gc8cr8ma2

Read more here: "Docs: Explain".

• Non-breaking: stable APIs
• Lazy evaluation: work is done only when needed
• Explainability: identifiers carry meaning
• Transparency: performance claims are backed by data

Think of TypeID as UUIDs + semantics + observability — without sacrificing speed

MIT