kazenai-finops

Stop your AI agents from burning your budget. Catch loops before they catch you.

Install: PyPI · kazenai-finops · Docs: docs.kazenai.com · Products: kazenai.com

Source: github.com/kazenai-ai/kazenai-finops-sdk

What is this?

kazenai-finops is the customer-facing Agent FinOps SDK. Install it, wrap your OpenAI or Anthropic client with monitor(), and get:

1. Pre-call budget deny — hard BudgetExceeded before a provider call
2. Soft trajectory pause — KazenCircuitBreaker when projected spend trips (alias: KazenBudgetExceeded)
3. Loop detection — stop repeated high-risk patterns before another LLM call
4. Optional FinOps / Lens evidence — canonical KazenEvent batches when ingest is configured

It depends on kazenai (the core engine) and kazen-event-schema. Most teams should install this package, not core directly.

Traditional tools:  LLM call → response → log cost → dashboard shows overspend
KazenAI:            Pre-flight check → BLOCKED → LLM call never made

Install

python -m pip install "kazenai-finops[openai]==1.1.1"
# Optional Anthropic path:
# python -m pip install "kazenai-finops[anthropic]==1.1.1"
# Both providers:
# python -m pip install "kazenai-finops[providers]==1.1.1"

Requires Python 3.10–3.12.

Provider extras pin the certified SDK ranges so a first-run install cannot silently resolve OpenAI 3.x or Anthropic 1.x. The Control 1.1.x certification covers those ranges. Later major versions are unverified until they are added to the supported matrix. Release 1.1.1 is a compatibility/distribution patch, not a new runtime capability line.

Quick start

from kazenai_finops import monitor, BudgetExceeded, KazenCircuitBreaker
import openai

client = openai.OpenAI()
monitored = monitor(
    client,
    agent_id="customer-support",
    # Optional ingest: set KAZENAI_FINOPS_API_KEY / KAZENAI_FINOPS_URL in the environment
    max_budget_usd=5.00,
    debug=True,
)

try:
    for i in range(100):
        monitored.chat.completions.create(
            model="gpt-4o-mini",
            messages=[{"role": "user", "content": f"Step {i}"}],
        )
except BudgetExceeded as e:
    print(f"Hard cap (before provider): {e}")
except KazenCircuitBreaker as e:
    print(f"Soft pause (after a completed call): {e}")

Call sites stay the same — monitor() wraps the client.

Supported today

Optional framework extras

pip install 'kazenai-finops[langgraph]'   # similarly: langchain, crewai, etc. — see pyproject.toml

from kazenai_finops.adapters.langchain import wrap_langchain_runnable
from kazenai_finops.adapters.langgraph import wrap_graph_invoke
from kazenai_finops.adapters.crewai import wrap_crew_kickoff

These helpers are for evaluation. For the supported control path, wrap the OpenAI/Anthropic client with monitor().

Streaming (Control-certified)

Streaming uses the same pre-dispatch reservation. At provider dispatch the reservation becomes provider_started. Settlement happens when the stream ends with authoritative usage; early close(), a provider error, or missing final usage becomes outcome_unknown and leaves shared reconciliation pending rather than recording an exact zero or releasing an already-started call.

from kazenai_finops import monitor, StreamCutoffError
import openai

monitored = monitor(
    openai.OpenAI(),
    agent_id="streaming-agent",
    max_budget_usd=1.00,
    stream_cutoff_usd=0.50,  # optional local cutoff (does not use /v1/budget/stream-tick)
)

try:
    stream = monitored.chat.completions.create(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": "Long answer please"}],
        stream=True,
    )
    with stream:
        for chunk in stream:
            ...
except StreamCutoffError as e:
    print(f"Stream cut at ${e.blocked_usd:.6f}")

The official lazy OpenAI helper is supported too. Dispatch occurs when its context manager is entered, and the nested SDK call is accounted for exactly once:

with monitored.chat.completions.stream(
    model="gpt-4o-mini",
    messages=[{"role": "user", "content": "Long answer please"}],
) as stream:
    for event in stream:
        ...

stream_enforcement=True is deprecated; prefer stream_cutoff_usd. The cutoff uses output observable at the client. It cannot see hidden reasoning or guarantee that the provider stopped generating or billing immediately after the close.

Local and shared enforcement

• Local hard cap works offline — no FinOps round-trip is required for the in-process client limit
• Low overhead — checks stay on the hot path
• Shared policies can fail closed — production/staging modes deny when a required shared reservation cannot be obtained

The local cap is not an account-wide provider billing limit. Multiple workers or services need the shared FinOps authority for a cross-process budget. An explicit development fail-open path retains only local safeguards.

How it relates to kazenai

from kazenai_finops import monitor re-exports the core engine so existing integrations keep working.

Maintainers: follow RELEASING.md. Core 1.1.1 must be available on public PyPI before this package's 1.1.1 CI, clean build and publication.

License

Apache License 2.0. See LICENSE and NOTICE.

Issues: https://github.com/kazenai-ai/kazenai-finops-sdk/issues · Products: kazenai.com · founder@kazenai.com