3tears-models

LangChain-native AI model factories for the 3tears framework. Build chat and embedding models from a single call, with capability metadata, circuit breakers, error translation, and usage tracking wired in.

pip install 3tears-models

What you get

Factories -- create_chat_model() and create_embedding_model() return standard LangChain BaseChatModel / Embeddings instances. No custom runtime protocols to learn.
Providers -- Anthropic, OpenAI, OpenRouter, VoyageAI, Whisper, and image backends (OpenAI Images, HuggingFace, A1111, ModelsLab, ComfyUI).
Capability registry -- get_capabilities(), register_capabilities(), and per-model overrides describe context windows, vision support, tool support, and tier.
Circuit breakers -- CircuitBreaker and CircuitBreakerRegistry trip on repeated provider failures and recover on a timer. The factory's default breaker is per provider and per API key, so one caller's failing key never fast-fails another key on the same provider.
Usage tracking -- UsageTracker.record() emits an OpenTelemetry span plus Prometheus instruments, with optional per-application audit and counter sinks.
Message hygiene -- preprocess_messages(), enforce_alternating_roles(), filter_invalid_tool_calls(), and streaming chunk helpers (parse_chunk, merge_chunks).

Quickstart

from threetears.models import create_chat_model, create_embedding_model

chat = create_chat_model("claude-sonnet-4-6")
reply = await chat.ainvoke("Summarize three-tier caching in one sentence.")

embedder = create_embedding_model("voyage-3")
vectors = await embedder.aembed_documents(["first doc", "second doc"])

Capabilities, circuit breakers, and usage tracking attach automatically. Reach for the registry and tracker APIs directly when you need to inspect or override them.

License

MIT. See LICENSE.

Metadata

Release files for 3tears-models 0.56.3

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for 3tears-models 0.56.3
File	Size	Uploaded
3tears_models-0.56.3.tar.gz	298.2 kB	Sep 28, 2026	Details

Built distribution (wheel)

Table of built distributions (wheels) for 3tears-models 0.56.3
File	Interpreter	ABI	Platform	Reset
3tears_models-0.56.3-py3-none-any.whl 167.2 kB Sep 28, 2026	Python 3	none	any	Details

Total release size: 465.3 kB

Release files / 3tears_models-0.56.3.tar.gz

Download URL	3tears_models-0.56.3.tar.gz
Size	298.2 kB
Tags	Source
SHA-256 checksum How to use checksums	`aed50126c8e807bed1de327470939c3ca4d3221cfaf32f77bed4457e68e424a1`
BLAKE2b-256 checksum How to use checksums	`510bbc9fad8f156f3cc9c8eafa807f524abfd8b86fdb537fd9d83440328081fc`
Upload date	Sep 28, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	Yes
Uploaded via	`twine/7.0.0 CPython/3.13.14`

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Identity

Publishing platform: GitHub Actions
Publishing repository: github.com/pacepace/3tears
Publishing commit: github.com/pacepace/3tears/tree/b96f0a59c244e28e69a51d6dda88854210b9e0f9

Workflow

Publishing configuration: github.com/pacepace/3tears/blob/b96f0a59c244e28e69a51d6dda88854210b9e0f9/.github/workflows/release.yml
Publishing logs: github.com/pacepace/3tears/actions/runs/36484056869/attempts/1

Artifact

Subject: 3tears_models-0.56.3.tar.gz
SHA-256 checksum: aed50126c8e807bed1de327470939c3ca4d3221cfaf32f77bed4457e68e424a1 Verifying attestations

Release files / 3tears_models-0.56.3-py3-none-any.whl

Download URL	3tears_models-0.56.3-py3-none-any.whl
Size	167.2 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`0265f0ee0fa8de47c05d668047d5ead2c1fa0158a6a89729679c3ae8cec7a25f`
BLAKE2b-256 checksum How to use checksums	`8d1c8d5334393ca1d6a0919703d2da18ad852c38489e884a395d3b7c1aa3a66b`
Upload date	Sep 28, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	Yes
Uploaded via	`twine/7.0.0 CPython/3.13.14`

Provenance

Identity

Publishing platform: GitHub Actions
Publishing repository: github.com/pacepace/3tears
Publishing commit: github.com/pacepace/3tears/tree/b96f0a59c244e28e69a51d6dda88854210b9e0f9

Workflow

Publishing configuration: github.com/pacepace/3tears/blob/b96f0a59c244e28e69a51d6dda88854210b9e0f9/.github/workflows/release.yml
Publishing logs: github.com/pacepace/3tears/actions/runs/36484056869/attempts/1

Artifact

Subject: 3tears_models-0.56.3-py3-none-any.whl
SHA-256 checksum: 0265f0ee0fa8de47c05d668047d5ead2c1fa0158a6a89729679c3ae8cec7a25f Verifying attestations