Cryptographic 563-leaf LLM rate oracle, context-cliff auditor, and prompt cache optimizer.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Check one LLM request's token counts against a pricing snapshot and return a deterministic decision: allow, pre-cliff warning, hard-cliff reroute, or budget-gate block, with estimated costs and a cheaper fallback model where one exists. Estimates only, not billing. Receives model name and token counts only, never prompt content.
Model ID from list_supported_models (gpt-6-1-sol, claude-haiku-5-5, claude-sonnet-5-5, claude-opus-5-5, gpt-6-luna)
Optional trace ID (1..64 alphanumeric/underscore/hyphen chars).
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
claude mcp add --transport http etuki-finops-mcp https://www.etuki.com/api/v1/mcp
Check one LLM request's token counts against a pricing snapshot and return a deterministic decision: allow, pre-cliff warning, hard-cliff reroute, or budget-gate block, with estimated costs and a cheaper fallback model where one exists. Estimates only, not billing. Receives model name and token counts only, never prompt content.
| Argument | Type | Description |
|---|---|---|
| model* | string | Model ID from list_supported_models (gpt-6-1-sol, claude-haiku-5-5, claude-sonnet-5-5, claude-opus-5-5, gpt-6-luna) |
| pass_rate | number | |
| request_id | string | Optional trace ID (1..64 alphanumeric/underscore/hyphen chars). |
| retry_rate | number | |
| prompt_tokens* | integer | |
| cache_hit_rate | number | |
| budget_gate_usd | number | |
| allow_async_batch | boolean | |
| completion_tokens* | integer | |
| routed_cache_hit_rate | number | |
| cold_start_cache_write | boolean |
{
"type": "object",
"required": [
"model",
"prompt_tokens",
"completion_tokens"
],
"properties": {
"model": {
"type": "string",
"description": "Model ID from list_supported_models (gpt-6-1-sol, claude-haiku-5-5, claude-sonnet-5-5, claude-opus-5-5, gpt-6-luna)"
},
"pass_rate": {
"type": "number",
"default": 0.96,
"maximum": 1,
"exclusiveMinimum": 0
},
"request_id": {
"type": "string",
"pattern": "^[a-zA-Z0-9_-]{1,64}$",
"maxLength": 64,
"description": "Optional trace ID (1..64 alphanumeric/underscore/hyphen chars)."
},
"retry_rate": {
"type": "number",
"default": 0.05,
"maximum": 1,
"minimum": 0
},
"prompt_tokens": {
"type": "integer",
"maximum": 9007199254740991,
"minimum": 0
},
"cache_hit_rate": {
"type": "number",
"default": 0.4,
"maximum": 1,
"minimum": 0
},
"budget_gate_usd": {
"type": "number",
"default": 0.06,
"minimum": 0
},
"allow_async_batch": {
"type": "boolean",
"default": true
},
"completion_tokens": {
"type": "integer",
"maximum": 9007199254740991,
"minimum": 0
},
"routed_cache_hit_rate": {
"type": "number",
"maximum": 1,
"minimum": 0
},
"cold_start_cache_write": {
"type": "boolean",
"default": false
}
},
"additionalProperties": false
}Retrieve the cryptographic snapshot metadata (Merkle root, RS256 public key URL, snapshot date, and 12/12 CI self-test reference) for the current pricing snapshot.
No arguments.
{
"type": "object",
"properties": {},
"additionalProperties": false
}Search and filter across all 563 Merkle-committed LLM API and cloud infrastructure entities in the Etuki catalog by keyword.
| Argument | Type | Description |
|---|---|---|
| query | string | Model or vendor keyword (e.g. gpt-6.1, claude, gemini, deepseek, hunyuan) |
{
"type": "object",
"properties": {
"query": {
"type": "string",
"description": "Model or vendor keyword (e.g. gpt-6.1, claude, gemini, deepseek, hunyuan)"
}
},
"additionalProperties": false
}Return the 5 deterministic cliff-engine models supported by check_context_cliff along with their input/output rates, cache rates, cliff token thresholds, soft-block thresholds, fallback targets, pricing_snapshot_id, and Merkle root.
No arguments.
{
"type": "object",
"properties": {},
"additionalProperties": false
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 1, 2026.