🤖 AI Tools
Langfuse Trace Score Map from Project Inventory (No Invented Latency)
Turn a Langfuse project inventory into a trace score map only. No invented latency, token costs, or API keys.
0Reviews
Prompt
Act as a Langfuse observability engineer who only uses a pasted project inventory. You write a trace score map the inventory already supports. You do not invent latency, token costs, API keys, or private trace IDs. This is not a live Langfuse ingest and not a model cost quote. You work only from Inputs. Do not invent stats, citations, quotes, URLs, names, IDs, or records that are not in Inputs. Inputs: - Langfuse inventory I lock (projects, traces, scores, stubs): [Inventory] - Langfuse / SDK version notes I lock: [Version] - Project or environment label I may quote (or UNKNOWN): [Project] - Score names already present (or UNKNOWN): [Scores] - Span names already present (or UNKNOWN): [Spans] - Model cues already present (or UNKNOWN): [Models] - Dataset labels already present (or UNKNOWN): [Datasets] - Words I must not use: [Banned] - What I must never invent (latency, token costs, API keys, private trace IDs): [Never] - Output format: [Format] - Language: [Lang] Generate: 1. Honesty ledger: Inventory nouns, Version, Project, Scores, Spans, Models, Datasets, Lang. Forbidden: invented latency, token costs, API keys, private trace IDs. Banner: not a live Langfuse ingest; not a model cost quote. 2. Score map table: one row per Inventory score stub. Missing notes write NOT IN INPUTS. Use Trace, Span, Score, Dataset, and Model language only when Inventory supports it. 3. Span set: only names in Spans. Unnamed spans stay NOT IN INPUTS. 4. Version lock: print Version. Refuse Langfuse features newer than Version if Version is named. 5. Refuse list: inventing latency, inventing token costs, inventing API keys, inventing private trace IDs. 6. Compliance pass: quote Banned and Never hits. Cut them. Format as Format. Constraints: - Map from Inventory only. No invented latency. Teach Langfuse trace scores, not a Helicone dashboard swap. - Honor Version. No emojis.
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
Inventory: Project Harbor Quay RAG as pasted; Score faithfulness_cleat as pasted; Score relevance_river as pasted; Span retrieve_mooring as pasted; Session UNKNOWN. Version: Langfuse 2.60 Python SDK as pasted (do not invent unreleased AI judge APIs) Project: Harbor Quay RAG as pasted Scores: faithfulness_cleat as pasted; relevance_river as pasted. toxicity UNKNOWN. Spans: retrieve_mooring as pasted; generate_answer as pasted. rerank UNKNOWN. Models: gpt-4o-mini cue as pasted. claude UNKNOWN. Datasets: Harbor Eval v1 as pasted. River Hard UNKNOWN. Banned: guaranteed latency, token invent, key invent, trace invent Never: invent latency, invent token costs, invent API keys, invent private trace IDs Format: ledger + table + spans + version + refuse + compliance Lang: English
Example Output
1. Ledger. Inventory: Project Harbor Quay RAG; Score faithfulness_cleat; Score relevance_river; Span retrieve_mooring; Session UNKNOWN. Version Langfuse 2.60 Python SDK. Project Harbor Quay RAG. Scores faithfulness_cleat, relevance_river. Spans retrieve_mooring, generate_answer. Models gpt-4o-mini cue. Datasets Harbor Eval v1. Lang English. Banner: not a live Langfuse ingest; not a model cost quote. Forbidden: invented latency, token costs, API keys, private trace IDs. 2. Score map. 1. faithfulness_cleat | project Harbor Quay RAG as pasted | span retrieve_mooring as pasted | dataset Harbor Eval v1 as pasted | latency NOT IN INPUTS | token cost NOT IN INPUTS. 2. relevance_river | project Harbor Quay RAG as pasted | span generate_answer as pasted | toxicity NOT IN INPUTS. 3. Session | UNKNOWN so write Session NOT IN INPUTS. No invented VALUES beyond Inventory. 3. Span set. retrieve_mooring, generate_answer as Spans. rerank UNKNOWN so write rerank NOT IN INPUTS. 4. Version lock. Langfuse 2.60 Python SDK as pasted. Unreleased AI judge APIs not used. claude Model UNKNOWN. 5. Refuse. Invented latency, token costs, API keys, private trace IDs: refused. 6. Compliance. Banned hits none. Format ledger+table+spans+version+refuse+compliance. Gaps: Session, toxicity score, rerank span, claude model, River Hard dataset. 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.