🤖 AI Tools
LiteLLM Model Router Checklist from Gateway Notes (No Invented Token Totals)
Compile a LiteLLM model-router checklist from pasted gateway notes only. No invented token totals, latency percentiles, or cost ranks. Not MLOps consulting.
0Reviews
Prompt
Act as a LiteLLM model-router coordinator who only uses pasted gateway notes. You compile a model-router checklist the notes already support. You do not invent token totals, latency percentiles, cost ranks, or uptime percentages. This is not a live LiteLLM sync, not Portkey gateway automation, and not MLOps consulting advice. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Gateway notes I lock (route stubs, model cues, fallback fragments): [GatewayNotes] - LiteLLM version or deployment notes I lock: [Version] - Deployment or project label I may quote (or UNKNOWN): [DeployLabel] - Route labels already present (or UNKNOWN): [RouteLabels] - Model cues already present (or UNKNOWN): [ModelCues] - Fallback cues already present (or UNKNOWN): [FallbackCues] - Budget or key cues already present (or UNKNOWN): [BudgetKeyCues] - Words I must not use: [Banned] - What I must never invent (token totals, latency percentiles, cost ranks, uptime percentages): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: GatewayNotes nouns, Version, DeployLabel, RouteLabels, ModelCues, FallbackCues, BudgetKeyCues, Lang. Banner: not MLOps consulting; not a live LiteLLM sync. Forbidden: invented token totals, latency percentiles, cost ranks, uptime percentages. 2. Model-router checklist: one checkbox row per RouteLabels entry. Attach only ModelCues and FallbackCues named beside that route in GatewayNotes. Missing model write NOT IN INPUTS. 3. Budget key sketch: for each BudgetKeyCues entry, list routes that name it. Do not invent a 2.4M token total if absent. 4. Fallback caution block: quote FallbackCues only. Provider SLA packs not in GatewayNotes stay NOT IN INPUTS. 5. Refuse list: inventing 2.4M token totals, inventing p95 180ms, inventing cost rank #1, inventing 99.99% uptime. 6. Compliance pass: quote Banned and Never hits. Cut them. Print route and model counts from GatewayNotes only. Format as Format. Constraints: - Model-router checklist from GatewayNotes only. No invented token totals. - Honor Version. No emojis. Not a live LiteLLM proxy. Not MLOps consulting advice.
Instructions
Replace every [bracket] with your details before running. Works on ChatGPT, Claude, and Gemini.
Generated Output
This image was generated using the prompt above.
Examples
Example Input
GatewayNotes: route label Harbor Chat as pasted model gpt-4o-mini as pasted fallback claude-haiku as pasted; route label Quay Embed as pasted model text-embedding-3-small as pasted. Token invent NONE. Latency invent NONE. Version: LiteLLM as pasted (do not invent unreleased AI cost coach) DeployLabel: Cedar Pier gateway as pasted RouteLabels: Harbor Chat as pasted; Quay Embed as pasted. Pier Batch UNKNOWN. ModelCues: model gpt-4o-mini as pasted for Harbor Chat; model text-embedding-3-small as pasted for Quay Embed. Vision model UNKNOWN. FallbackCues: fallback claude-haiku as pasted for Harbor Chat; Quay Embed fallback NOT IN INPUTS. BudgetKeyCues: budget key HarborProd as pasted for Harbor Chat; Quay Embed key NOT IN INPUTS. Banned: 2.4M tokens, p95 180ms, guaranteed 99.99% uptime Never: invent token totals, invent latency percentiles, invent cost ranks, invent uptime percentages Format: ledger + model-router checklist + budget key sketch + fallback caution + refuse + compliance Lang: English
Example Output
1. Ledger. GatewayNotes: route label Harbor Chat + model gpt-4o-mini + fallback claude-haiku; route label Quay Embed + model text-embedding-3-small. Token invent NONE. Latency invent NONE. Version LiteLLM. DeployLabel Cedar Pier gateway. RouteLabels Harbor Chat; Quay Embed. ModelCues gpt-4o-mini for Harbor Chat; text-embedding-3-small for Quay Embed. FallbackCues claude-haiku for Harbor Chat; Quay Embed fallback NOT IN INPUTS. BudgetKeyCues HarborProd for Harbor Chat; Quay Embed key NOT IN INPUTS. Lang English. Banner: not MLOps consulting; not a live LiteLLM sync. Forbidden: invented token totals, latency percentiles, cost ranks, uptime percentages, 2.4M tokens, p95 180ms, guaranteed 99.99% uptime. 2. Model-router checklist. [ ] Harbor Chat | model gpt-4o-mini | fallback claude-haiku as pasted. [ ] Quay Embed | model text-embedding-3-small | fallback NOT IN INPUTS. Pier Batch not attached. Vision model not added. 3. Budget key sketch. budget key HarborProd | route Harbor Chat as pasted. Quay Embed key | NOT IN INPUTS. Token total NOT IN INPUTS so do not invent 2.4M tokens. Second HarborProd cue not invented. 4. Fallback caution. fallback claude-haiku as pasted for Harbor Chat. Quay Embed fallback NOT IN INPUTS. Provider SLA pack NOT IN INPUTS. Do not invent uptime packs. 5. Refuse. 2.4M token totals: refused. p95 180ms: refused. cost rank #1: refused. 99.99% uptime: refused. Unreleased AI cost coach: refused. 6. Compliance. Banned hits none. Routes 2. Models 2. Format ledger+model-router checklist+budget key sketch+fallback caution+refuse+compliance. Gaps: Quay Embed fallback, Quay Embed key, Pier Batch, Vision model, token total. Missing-data policy: if a field was blank, write NOT IN INPUTS rather than guessing. Lock any tool version named in Inputs; if unnamed, write unknown. No invented testimonials, star ratings, or press logos. If legal, clinical, insurance, HR, education-plan, or veterinary content appears, add a one-line not-advice and de-identify banner. Quote banned-word hits and cut them. End with a gaps list of five bullets the user still owes you. Character and byte caps in the job are hard; print counts when relevant. Refuse to backfill DOIs, exam dumps, PHI, PII, or compensation promises not in Inputs.