Skip to content

Glossary

TandemApp's terms. The engine's (anchor, round, StairCut and the rest) are in TandemEngine's glossary.

  • adapter The proxy's code for one backend kind: how it relays a route, where it reads cached and reasoning tokens, how it switches thinking. proxy/adapters/.
  • backend An engine or a cloud API behind the proxy: another TandemEngine, vLLM, SGLang, any OpenAI-compatible or Messages API server, OpenAI, Anthropic, DeepSeek. Each has a kind, a URL, models, an optional key, and switches (backends.md).
  • feature group One switchable part of the harness in front of a backend: class, effort, packages, side, loops, e1, tools, compaction, states, memory.
  • host collector The code that reads a machine's memory, CPU, GPUs, disks and the process behind each backend's port; tandem-app hostcollect on a remote machine.
  • L1, L2 The two layers of engine figures behind the proxy: L1 is the engine's /metrics page, L2 the plugin's per-request numbers.
  • preset A provider the Add backend form offers (OpenAI, OpenRouter, Z.ai, Ollama and the rest): its base URL for each protocol, its kind and how it takes a key. proxy/presets.py.
  • protocol What a backend or a client speaks: openai (chat completions, completions, embeddings, Responses) or anthropic (the Messages API).
  • plugin A small package inside vLLM or SGLang that posts each request's own times to the proxy.
  • proxy The router of requests by model id to backends, with one ledger and one dashboard: tandem-app, the controller (until engine API 1.8 also the engine's server with --backends).