Hyde Campfire
Campfire learns your enterprise’s AI work, routes every task by the metrics you set, and turns the repeatable ones into models you own.
Plugs into your stack.
- Your cloud: AWS, Azure, Google Cloud, Together, Fireworks, Baseten, vLLM on-prem
- Your traces: Langfuse, LangSmith, Datadog, Braintrust, OpenTelemetry, Snowflake, Databricks
- Your tools: Microsoft Copilot, Claude Code, Cursor, Codex, GitHub Copilot, Palantir AIP, LangChain, In-house apps
Your cloudYour tracesYour tools
AWS
Azure
Google Cloud
Together
Fireworks
Baseten
vLLM on-prem
AWS
Azure
Google Cloud
Together
Fireworks
Baseten
vLLM on-prem
AWS
Azure
Google Cloud
Together
Fireworks
Baseten
vLLM on-prem
Langfuse
LangSmith
Datadog
Braintrust
OpenTelemetry
Snowflake
Databricks
Langfuse
LangSmith
Datadog
Braintrust
OpenTelemetry
Snowflake
Databricks
Langfuse
LangSmith
Datadog
Braintrust
OpenTelemetry
Snowflake
Databricks
Microsoft Copilot
Claude Code
Cursor
Codex
GitHub Copilot
Palantir AIP
LangChain
In-house apps
Microsoft Copilot
Claude Code
Cursor
Codex
GitHub Copilot
Palantir AIP
LangChain
In-house apps
Microsoft Copilot
Claude Code
Cursor
Codex
GitHub Copilot
Palantir AIP
LangChain
In-house apps Your AI workload. Your model strategy.
Each trace adds task, model, spend and quality to your taskonomy.

Evals from your real tasks, not public leaderboards.

Frontier, open and sovereign models scored on your bar.

Routing improves as work and models change. Repeat.

Runs in your VPC, with PII redacted and every route logged.