forgesight 0.1.0

Vendor-neutral, OpenTelemetry-first telemetry for AI agents: traces, cost, budgets, and a tamper-evident audit trail — ship to any backend, no agent-code changes.

ForgeSight

Instrument any AI agent in a few lines — then ship traces, cost, metrics, evals, budgets, and a tamper-evident audit trail to any backend by changing one line of config. OpenTelemetry-first. Vendor-neutral. Never an agent-code change.

▶️ See the 30-second demo — instrument an agent and watch the trace, cost, and a verified tamper-evident audit trail appear.

forgesight is the batteries-included facade — the package most people install.

pip install "forgesight[otel]"

import forgesight
from forgesight import telemetry

forgesight.configure(exporters=["otel"])          # pick a backend by name — that's it

with telemetry.agent_run("pr-reviewer", version="2.1.0", metadata={"team": "platform"}) as run:
    with run.llm_call("anthropic", "claude-sonnet-4-5") as call:
        resp = await client.messages.create(...)
        call.record_usage(input=resp.usage.input_tokens, output=resp.usage.output_tokens)
    with run.tool_call("github_get_diff"):
        diff = gh.get_diff(pr)

Cost, token usage, trace nesting, metrics, and multi-backend fan-out come for free. Swap otel → langfuse, datadog, clickhouse, prometheus — the agent code never moves.

What you get

• 🔌 Vendor-neutral. The core depends on no backend or model-provider SDK. Backends are packages you install and select by config.
• 📐 OpenTelemetry-first. GenAI semantic conventions, so any OTLP backend works with no dedicated package.
• 💰 Cost built in. Token usage → USD via a pluggable pricing table — the same number everywhere — plus live attribution by team and pre-call budget projection.
• 🚦 Non-blocking & fault-isolated. Export runs on a background worker; a backend outage never breaks a run.
• 🔒 Secure by default. Content capture is opt-in; redaction runs before export.
• 🧾 Accountability. A tamper-evident, hash-chained audit trail with a compliance query/export surface.

Install backends & integrations as extras

pip install "forgesight[otel]"                      # one backend (OTLP → any OTel platform)
pip install "forgesight[otel,langfuse,datadog]"     # several
pip install "forgesight[all]"                       # everything except the heavy CrewAI tree

Extras: otel · prometheus · langfuse · clickhouse · datadog · mcp · fastapi · github · governance · eval · registry · audit · adapters-langgraph · adapters-crewai. Installing a package enables a backend; config selects it.

Docs

• README & guides — full docs, playbooks, runbooks
• Quick start · Run locally with Docker · Examples

License

Apache-2.0