langchain-finic 0.1.0a4__tar.gz

langchain_finic-0.1.0a4/.gitignore

@@ -0,0 +1,59 @@
+# Python bytecode and compiled extensions
+__pycache__/
+*.py[cod]
+*.so
+
+# Virtual environments
+.venv/
+venv/
+env/
+
+# Packaging and build output
+build/
+dist/
+*.egg-info/
+.eggs/
+
+# Test, coverage, and tooling caches
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.pyright/
+.tox/
+.nox/
+.coverage
+.coverage.*
+coverage.xml
+htmlcov/
+.cache/
+.ipynb_checkpoints/
+
+# Local secrets; keep templates trackable
+.env
+.env.*
+!.env.example
+!.env.*.example
+!.env.sample
+!.env.template
+
+# Logs and editor files
+*.log
+.idea/
+.vscode/
+*.swp
+*.swo
+*~
+
+# Operating system metadata
+.DS_Store
+Thumbs.db
+Desktop.ini
+
+# Local case reports and cached model responses
+artifacts/
+
+# Downloaded third-party India benchmark sources; provenance and checksums are tracked.
+data/india/raw/
+
+# Generated documentation
+site/

langchain_finic-0.1.0a4/CHANGELOG.md

@@ -0,0 +1,15 @@
+# Changelog
+
+## 0.1.0a4
+
+Initial public alpha. See [GitHub Releases](https://github.com/supreplabs/langchain-finic/releases)
+for publication dates and distribution checksums.
+
+- Typed `Noul`, `Choice`, `MultiChoice`, and `Score` decisions with response validation.
+- LangChain Runnable composition, tools, sync/async calls, batching, and tracing metadata.
+- Optional model-routing and tool-screening middleware for LangChain agents.
+- Configurable model name, transport injection, bounded retries, and client lifecycle management.
+- MIT license, typed distributions, usage documentation, and API reference.
+
+Configuration uses `FINIC_*` environment variables. Earlier unpublished prototypes
+used `GRAYSON_*` aliases; those aliases are not supported by this release.

langchain_finic-0.1.0a4/CODE_OF_CONDUCT.md

@@ -0,0 +1,14 @@
+# Community conduct
+
+Be respectful, constructive, and welcoming to people of all backgrounds and levels
+of experience. Focus feedback on the work. Harassment, discrimination, threats,
+personal attacks, and disclosure of private information are not acceptable.
+
+Maintainers may edit or remove inappropriate contributions and restrict participation
+when needed. Enforcement should be proportionate and explained to the affected person
+when possible. Apply these expectations consistently in issues, reviews, discussions,
+and other project spaces.
+
+Report conduct concerns privately to the Suprep Labs organization maintainers through
+an available private contact on their GitHub profiles. Do not post private details in
+public issues. Security vulnerabilities use the separate reporting path in SECURITY.md.

langchain_finic-0.1.0a4/CONTRIBUTING.md

@@ -0,0 +1,54 @@
+# Contributing
+
+Use Python 3.11 or later and [uv](https://docs.astral.sh/uv/).
+
+```bash
+git clone https://github.com/supreplabs/langchain-finic.git
+cd langchain-finic
+make install
+make check
+```
+
+Keep changes focused, describe the behavior and validation in your pull request,
+and add regression tests for bugs. Public imports and serialized class identifiers
+are compatibility commitments, including during the alpha period. Discuss breaking
+changes in an issue before implementing them. Use conventional commit titles.
+
+`tests/unit_tests` includes mocked framework composition tests and LangChain's
+standard tool tests. Network access is disabled. Tests must not load `.env` or
+require credentials. `langchain-tests` is pinned because its test contracts can
+change independently of this package. Update it deliberately with the lockfile.
+
+`make integration_test` makes two paid API requests (sync and async), with one
+attempt each, using synthetic evidence. It requires an exported `FINIC_API_KEY`;
+never put keys, production inputs, or raw responses in issues or fixtures.
+These tests check response contracts, not deterministic probability values.
+
+Use `make format` for formatting and `uv lock` after dependency changes.
+Documentation lives in `docs/`; public API pages are generated from docstrings.
+See [release instructions](RELEASING.md) for distribution and publishing.
+
+By contributing, you agree to license your contributions under this repository's
+MIT license. See [community conduct](CODE_OF_CONDUCT.md) and
+[security reporting](SECURITY.md).
+
+The minimum-dependency CI job uses `langchain-tests==1.0.0`, which supports
+`langchain-core==1.0.0`; current-dependency jobs use `langchain-tests==1.1.9`.
+The latter requires core >=1.4 and cannot validate the historical minimum.
+
+## Documentation standards
+
+Keep the user site focused on installation, supported interfaces, configuration,
+examples, and API behavior. Release procedures and directory-listing instructions
+belong in the repository's maintainer guides, outside `docs/`.
+
+Ground capability claims in the public API and tests. Do not publish internal
+research results, account-specific usage, or unverified accuracy/calibration claims.
+Use fictional examples. Define inputs explicitly or label examples as reusable
+functions with caller-supplied parameters. Python snippets in the README and user
+guides are executed offline by `test_documentation.py`; keep them runnable in the
+order shown. Prefer links over duplicated explanations and check relative links.
+
+Follow [LangChain's documentation quality guidance](https://docs.langchain.com/oss/python/contributing/documentation#quality-standards).
+Its Mintlify-specific syntax applies to LangChain-hosted pages; this repository's
+site uses MkDocs and generated Python API references.

langchain_finic-0.1.0a4/LANGCHAIN_LISTING.md

@@ -0,0 +1,22 @@
+# LangChain directory listing
+
+Submit only after PyPI publication. This document prepares the listing; it does not
+claim acceptance, partnership, or endorsement.
+
+Use the [LangChain integration listing process](https://docs.langchain.com/oss/python/contributing/publish-langchain).
+New packages normally receive an external-docs listing rather than a hosted guide.
+Do not open a manual documentation PR unless a maintainer asks.
+
+| Field | Value |
+| --- | --- |
+| Display name | Finic |
+| Language | Python |
+| Package | langchain-finic |
+| Component | Tools; middleware (select the form's corresponding category) |
+| Documentation | https://supreplabs.github.io/langchain-finic/ |
+| Repository | https://github.com/supreplabs/langchain-finic |
+| Description | Typed Finic Grayson decisions as LangChain tools and Runnables, with optional model routing and tool-screening middleware. |
+
+`FinicClassifier.as_tool()` produces a standard LangChain tool. Middleware classes
+extend `AgentMiddleware`. The integration is not a chat model; do not advertise
+chat generation, token streaming, or provider-native tool calling.

langchain_finic-0.1.0a4/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Suprep Labs
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

langchain_finic-0.1.0a4/Makefile

@@ -0,0 +1,21 @@
+.PHONY: install test integration_test lint format type_check docs build check
+install:
+	uv sync --locked --group dev --extra middleware
+test:
+	uv run --locked --group test pytest tests/unit_tests -q
+integration_test:
+	FINIC_LIVE_TESTS=1 uv run --locked --group test pytest -o addopts= tests/integration_tests -m live -q
+lint:
+	uv run --locked --group lint ruff check src tests examples scripts
+	uv run --locked --group lint ruff format --check src tests examples scripts
+format:
+	uv run --locked --group lint ruff format src tests examples scripts
+type_check:
+	uv run --locked --group lint --extra middleware mypy
+docs:
+	uv run --locked --group docs --extra middleware mkdocs build --strict
+build:
+	uv build
+	uv run --locked --group build twine check dist/*
+	uv run --locked python scripts/check_dist.py
+check: lint type_check test docs build

langchain_finic-0.1.0a4/PKG-INFO

@@ -0,0 +1,142 @@
+Metadata-Version: 2.5
+Name: langchain-finic
+Version: 0.1.0a4
+Summary: Community LangChain tools, Runnables, and middleware for Finic Grayson decisions
+Project-URL: Homepage, https://github.com/supreplabs/langchain-finic
+Project-URL: Documentation, https://supreplabs.github.io/langchain-finic/
+Project-URL: Repository, https://github.com/supreplabs/langchain-finic
+Project-URL: Issues, https://github.com/supreplabs/langchain-finic/issues
+Project-URL: Changelog, https://github.com/supreplabs/langchain-finic/blob/main/CHANGELOG.md
+Author: Suprep Labs
+Maintainer: Suprep Labs
+License-Expression: MIT
+License-File: LICENSE
+Keywords: agents,finic,grayson,langchain,langgraph
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Classifier: Typing :: Typed
+Requires-Python: >=3.11
+Requires-Dist: httpx<1,>=0.28.1
+Requires-Dist: langchain-core<2,>=1
+Requires-Dist: langsmith<1,>=0.3
+Requires-Dist: pydantic<3,>=2.12
+Provides-Extra: middleware
+Requires-Dist: langchain<2,>=1; extra == 'middleware'
+Requires-Dist: langgraph<2,>=1; extra == 'middleware'
+Description-Content-Type: text/markdown
+
+# langchain-finic
+
+Community-maintained LangChain integration for **Finic Grayson**, by Suprep Labs.
+Typed decisions, LCEL composition, LangGraph nodes, structured tools, and optional
+agent middleware. This alpha is not endorsed by Finic or LangChain.
+
+[Documentation](https://supreplabs.github.io/langchain-finic/) ·
+[Source](https://github.com/supreplabs/langchain-finic) ·
+[PyPI](https://pypi.org/project/langchain-finic/) ·
+[MIT license](https://github.com/supreplabs/langchain-finic/blob/main/LICENSE)
+
+## Install
+
+Python 3.11 or newer. Install the alpha release:
+
+```bash
+pip install langchain-finic==0.1.0a4
+export FINIC_API_KEY=your_key
+# Optional; defaults to grayson-1:
+export FINIC_MODEL=grayson-1
+```
+
+## Make a typed decision
+
+```python
+from langchain_finic import Choice, DecisionRequest, FinicClassifier, Noul
+
+request = DecisionRequest(
+    state={"events": ["New device", "Transfer to an unfamiliar recipient"]},
+    questions={
+        "takeover": Noul(instructions="Does this evidence suggest account takeover?"),
+        "route": Choice(
+            instructions="Choose a review queue.",
+            criteria={"standard": "Routine review", "urgent": "Urgent investigation"},
+        ),
+    },
+)
+
+with FinicClassifier() as classifier:
+    decision = classifier.invoke(request)
+    print(decision.nouls["takeover"].noul)
+    print(decision.choices["route"].choice)
+    print(decision.request_id)
+```
+
+`Noul`, `Choice`, `MultiChoice`, and `Score` have typed answers. Responses retain
+probabilities, usage, and request metadata. Finic evaluates decisions; it does not
+generate chat messages or text tokens.
+
+## Compose with LangChain and LangGraph
+
+```python
+# request is defined above.
+with FinicClassifier() as classifier:
+    route_chain = classifier | (lambda result: result.choices["route"].choice)
+    route = route_chain.invoke(request, config={"tags": ["risk-review"]})
+    tool = classifier.as_tool()
+    print(tool.name)  # finic_decide; pass it to create_agent or ToolNode
+```
+
+```python
+import asyncio
+
+async def review():
+    async with FinicClassifier() as classifier:
+        return await classifier.abatch([request], config={"max_concurrency": 1})
+
+results = asyncio.run(review())
+```
+
+Batching makes individual API requests; concurrency is controlled by Runnable config.
+Streaming yields one complete decision. A chat-provider model can call the Finic tool
+through `create_agent`; the Finic integration itself is a decision Runnable.
+
+## Optional middleware
+
+```bash
+pip install 'langchain-finic[middleware]==0.1.0a4'
+```
+
+`ModelRouterMiddleware` selects among host chat models. `AutoModeMiddleware` screens
+specified tools before execution. See the complete
+[middleware example](https://github.com/supreplabs/langchain-finic/blob/main/examples/middleware_agent.py)
+and [configuration guide](https://supreplabs.github.io/langchain-finic/usage/).
+Tool screening is probabilistic; application authorization remains necessary.
+
+## Configuration and data
+
+Use `FINIC_API_KEY`, `FINIC_MODEL`, `FINIC_TIMEOUT_SECONDS`, and `FINIC_MAX_ATTEMPTS`,
+or constructor options. The library does not load `.env`. The host controls optional
+LangSmith tracing, which can record request data and responses. Default timeout is
+60 seconds, with up to three attempts; retries can incur additional charges.
+Use context managers to reuse and close owned HTTP clients. Injected clients remain
+caller-owned. Read the [usage guide](https://supreplabs.github.io/langchain-finic/usage/)
+for precedence, serialization, resource lifetimes, and error handling.
+
+## Development
+
+```bash
+uv sync --locked --group dev --extra middleware
+make check
+```
+
+Default tests use mocked HTTP with network disabled. Opt-in live tests spend Finic
+credits. See [contributing](https://github.com/supreplabs/langchain-finic/blob/main/CONTRIBUTING.md)
+and [security reporting](https://github.com/supreplabs/langchain-finic/blob/main/SECURITY.md).
+
+References: [Finic API](https://docs.finic.ai/reference/decide),
+[LangChain integration guidelines](https://docs.langchain.com/oss/python/contributing/integrations-langchain).

langchain_finic-0.1.0a4/README.md

@@ -0,0 +1,109 @@
+# langchain-finic
+
+Community-maintained LangChain integration for **Finic Grayson**, by Suprep Labs.
+Typed decisions, LCEL composition, LangGraph nodes, structured tools, and optional
+agent middleware. This alpha is not endorsed by Finic or LangChain.
+
+[Documentation](https://supreplabs.github.io/langchain-finic/) ·
+[Source](https://github.com/supreplabs/langchain-finic) ·
+[PyPI](https://pypi.org/project/langchain-finic/) ·
+[MIT license](https://github.com/supreplabs/langchain-finic/blob/main/LICENSE)
+
+## Install
+
+Python 3.11 or newer. Install the alpha release:
+
+```bash
+pip install langchain-finic==0.1.0a4
+export FINIC_API_KEY=your_key
+# Optional; defaults to grayson-1:
+export FINIC_MODEL=grayson-1
+```
+
+## Make a typed decision
+
+```python
+from langchain_finic import Choice, DecisionRequest, FinicClassifier, Noul
+
+request = DecisionRequest(
+    state={"events": ["New device", "Transfer to an unfamiliar recipient"]},
+    questions={
+        "takeover": Noul(instructions="Does this evidence suggest account takeover?"),
+        "route": Choice(
+            instructions="Choose a review queue.",
+            criteria={"standard": "Routine review", "urgent": "Urgent investigation"},
+        ),
+    },
+)
+
+with FinicClassifier() as classifier:
+    decision = classifier.invoke(request)
+    print(decision.nouls["takeover"].noul)
+    print(decision.choices["route"].choice)
+    print(decision.request_id)
+```
+
+`Noul`, `Choice`, `MultiChoice`, and `Score` have typed answers. Responses retain
+probabilities, usage, and request metadata. Finic evaluates decisions; it does not
+generate chat messages or text tokens.
+
+## Compose with LangChain and LangGraph
+
+```python
+# request is defined above.
+with FinicClassifier() as classifier:
+    route_chain = classifier | (lambda result: result.choices["route"].choice)
+    route = route_chain.invoke(request, config={"tags": ["risk-review"]})
+    tool = classifier.as_tool()
+    print(tool.name)  # finic_decide; pass it to create_agent or ToolNode
+```
+
+```python
+import asyncio
+
+async def review():
+    async with FinicClassifier() as classifier:
+        return await classifier.abatch([request], config={"max_concurrency": 1})
+
+results = asyncio.run(review())
+```
+
+Batching makes individual API requests; concurrency is controlled by Runnable config.
+Streaming yields one complete decision. A chat-provider model can call the Finic tool
+through `create_agent`; the Finic integration itself is a decision Runnable.
+
+## Optional middleware
+
+```bash
+pip install 'langchain-finic[middleware]==0.1.0a4'
+```
+
+`ModelRouterMiddleware` selects among host chat models. `AutoModeMiddleware` screens
+specified tools before execution. See the complete
+[middleware example](https://github.com/supreplabs/langchain-finic/blob/main/examples/middleware_agent.py)
+and [configuration guide](https://supreplabs.github.io/langchain-finic/usage/).
+Tool screening is probabilistic; application authorization remains necessary.
+
+## Configuration and data
+
+Use `FINIC_API_KEY`, `FINIC_MODEL`, `FINIC_TIMEOUT_SECONDS`, and `FINIC_MAX_ATTEMPTS`,
+or constructor options. The library does not load `.env`. The host controls optional
+LangSmith tracing, which can record request data and responses. Default timeout is
+60 seconds, with up to three attempts; retries can incur additional charges.
+Use context managers to reuse and close owned HTTP clients. Injected clients remain
+caller-owned. Read the [usage guide](https://supreplabs.github.io/langchain-finic/usage/)
+for precedence, serialization, resource lifetimes, and error handling.
+
+## Development
+
+```bash
+uv sync --locked --group dev --extra middleware
+make check
+```
+
+Default tests use mocked HTTP with network disabled. Opt-in live tests spend Finic
+credits. See [contributing](https://github.com/supreplabs/langchain-finic/blob/main/CONTRIBUTING.md)
+and [security reporting](https://github.com/supreplabs/langchain-finic/blob/main/SECURITY.md).
+
+References: [Finic API](https://docs.finic.ai/reference/decide),
+[LangChain integration guidelines](https://docs.langchain.com/oss/python/contributing/integrations-langchain).

langchain_finic-0.1.0a4/RELEASING.md

@@ -0,0 +1,96 @@
+# Releasing
+
+Release target: **0.1.0a4**, repository **supreplabs/langchain-finic**.
+The package remains alpha. A release is complete only when PyPI installation is verified.
+
+## One-time account setup
+
+Create separate accounts on [PyPI](https://pypi.org/account/register/) and
+[TestPyPI](https://test.pypi.org/account/register/). Verify each email address and
+configure two-factor authentication (security key/passkey or authenticator), saving
+recovery codes privately. Do not share passwords, recovery codes, or API keys here.
+
+Register a **pending publisher** on each site's account publishing page:
+[PyPI](https://pypi.org/manage/account/publishing/) and
+[TestPyPI](https://test.pypi.org/manage/account/publishing/).
+Choose GitHub Actions and enter these exact values:
+
+| Field | TestPyPI | PyPI |
+| --- | --- | --- |
+| PyPI project name | `langchain-finic` | `langchain-finic` |
+| GitHub owner | `supreplabs` | `supreplabs` |
+| Repository name | `langchain-finic` | `langchain-finic` |
+| Workflow filename | `release.yml` | `release.yml` |
+| Environment name | `testpypi` | `pypi` |
+
+Pending publishers create the project on first successful upload; they do not reserve
+its name. If a project already exists, verify ownership and configure its publisher
+through the project settings. Never rename or upload to another project automatically.
+No PyPI API token or `PYPI_TOKEN` GitHub secret is needed.
+
+Reference: [PyPI pending publishers](https://docs.pypi.org/trusted-publishers/creating-a-project-through-oidc/).
+
+## Repository configuration
+
+GitHub environments `testpypi` and `pypi` permit deployments only from tags matching
+`v*`. Enable GitHub Pages with GitHub Actions as the source, private vulnerability
+reporting, and required CI checks on the default branch. CI and documentation builds
+need no API credentials. Publishing jobs alone receive OIDC permission.
+
+## Validate and release
+
+1. Update `pyproject.toml` version and changelog, run `uv lock`, and commit the release.
+2. Run `make install` and `make check`. Confirm all compatibility CI jobs pass.
+3. Inspect wheel/sdist contents, generated documentation, and `scripts/check_dist.py`
+   hash output. No benchmark datasets, research code, credentials, or case artifacts
+   belong in this repository or its distributions.
+4. Confirm both pending publishers are configured and the package name is still
+   available to this maintainer. Create and push the matching version tag:
+
+```bash
+git tag -a v0.1.0a4 -m 'Release 0.1.0a4'
+git push origin v0.1.0a4
+```
+
+The release workflow reruns CI, checks the tag against metadata, builds the release
+wheel and sdist once, and uploads immutable workflow artifacts with SHA-256 hashes.
+It publishes to TestPyPI, retrieves the exact files and checks their hashes, then
+installs the wheel in a fresh environment. Dependencies are resolved from PyPI only;
+TestPyPI is not used as an extra dependency index.
+
+The same artifacts are then published to PyPI with attestations, downloaded and
+hash-verified, and installed by exact public version. A GitHub prerelease is created
+only after installation succeeds. All action references are pinned to commit SHAs.
+
+If publishing succeeds but a later job fails, rerun failed jobs in the **same run**
+to preserve the exact build artifacts. Existing files are skipped during upload,
+but registry hash verification still requires an exact match. Do not rebuild and
+replace an already published version; use a new version for changed artifacts.
+`workflow_dispatch` must be dispatched on the version tag, not `main`.
+
+```bash
+pip install langchain-finic==0.1.0a4
+pip install 'langchain-finic[middleware]==0.1.0a4'
+```
+
+## Live smoke checks
+
+Default tests and installation checks use mock HTTP and spend no credits. To make
+two explicit live requests with fictional data:
+
+```bash
+export FINIC_API_KEY=your_key
+make integration_test
+```
+
+Do not commit the key or record production inputs. Tests verify the response contract,
+not an exact probability. No live calls run in pull requests or release CI.
+
+## After publication
+
+Use [listing metadata](LANGCHAIN_LISTING.md) for a LangChain integration listing request; do not
+claim LangChain or Finic endorsement. Obtain maintainer authorization before submitting
+an issue on their behalf.
+
+If a release is broken, publish a corrected version and consider yanking the bad
+version with a clear reason. Do not delete published files or move release tags.

langchain_finic-0.1.0a4/SECURITY.md

@@ -0,0 +1,14 @@
+# Security
+
+Security fixes target the latest published alpha. Older pre-releases are not maintained.
+
+Use [GitHub private vulnerability reporting](https://github.com/supreplabs/langchain-finic/security/advisories/new).
+Include affected versions, reproduction steps, impact, and a minimal example with fake
+credentials and synthetic data. Do not disclose vulnerabilities through public issues.
+
+The library sends request state and questions to Finic. Host-enabled LangSmith tracing
+can also record inputs and responses. The application controls consent, access checks,
+redaction, tracing, retention, and environment loading. The configured API key is excluded from
+configuration representations and serialized values. Request state and caller metadata
+are not automatically redacted. Load serialized objects only from
+trusted sources. Tool screening is probabilistic and does not grant authorization.

langchain_finic-0.1.0a4/docs/api.md

@@ -0,0 +1,43 @@
+# API reference
+
+## Classifier
+
+::: langchain_finic.classifier.FinicClassifier
+    options:
+      members: [model, api_key, timeout, max_attempts, client, http_client, http_async_client, transport, async_transport, invoke, ainvoke, as_tool, close, aclose, to_json]
+
+Inherited Runnable methods such as `batch`, `abatch`, `bind`, and `with_config`
+follow the [LangChain Runnable interface](https://reference.langchain.com/python/langchain-core/runnables).
+See the [usage guide](usage.md) for resource lifetime and invocation behavior.
+
+## HTTP client and errors
+
+::: langchain_finic.client.FinicClient
+    options:
+      members: [__init__, decide, adecide, close, aclose]
+
+::: langchain_finic.client.FinicError
+
+## Request and response types
+
+::: langchain_finic.models
+    options:
+      members: [DecisionRequest, DecisionResponse, Noul, Choice, MultiChoice, Score, NoulAnswer, ChoiceAnswer, MultiChoiceAnswer, ScoreAnswer]
+
+`Noul` and `Score` are public aliases. Their fields are defined by the following schemas:
+
+::: langchain_finic.models.NoulQuestion
+
+::: langchain_finic.models.ScoreQuestion
+
+`Choice` and `MultiChoice` share the fields defined by `ChoiceQuestion`:
+
+::: langchain_finic.models.ChoiceQuestion
+
+## Optional middleware
+
+Requires `langchain-finic[middleware]`.
+
+::: langchain_finic.middleware
+    options:
+      members: [ModelChoice, ModelRouterMiddleware, AutoModeMiddleware, ModelRouterState]