LLM Pricing

Measured conclusion: Deep-tested. L2 outcome: tools_ok.

Real-time, drift-verified LLM pricing across 26 providers with cost+quality model routing.

Add to your agent

Copy this into your MCP client config:

{
  "mcpServers": {
    "LLM Pricing": {
      "url": "https://llm-pricing-api.fly.dev/mcp",
      "transport": "streamable-http"
    }
  }
}

Measured tools (29)

• get_all_pricing

  Get current pricing for all LLM models from all providers

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• estimate_cost

  Estimate the cost of using a specific LLM model

  Parameters:

  • model_name (required) — string — Name of the LLM model (e.g., 'gpt-4', 'claude-3-opus')
  • input_tokens (required) — integer — Number of input tokens
  • output_tokens (required) — integer — Number of output tokens

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• compare_costs

  Compare costs for multiple LLM models

  Parameters:

  • model_names (required) — array — List of model names to compare
  • input_tokens (required) — integer — Number of input tokens
  • output_tokens (required) — integer — Number of output tokens

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_performance_metrics

  Get performance metrics (throughput, latency, context window) for models

  Parameters:

  • provider — string — Optional provider filter (e.g., 'openai', 'anthropic')
  • include_cost — boolean — Include cost metrics in response (default: true)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_use_cases

  Get recommended use cases and strengths for LLM models

  Parameters:

  • provider — string — Optional provider filter

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_telemetry

  Get MCP server telemetry and usage statistics including tool usage, response times, and error rates

  Parameters:

  • limit — integer — Maximum items per category (default: 10, max: 50)
  • include_details — boolean — Include detailed statistics (default: true)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_pricing_history

  Query historical pricing snapshots recorded over time. Use this to look up past prices for a model or provider, or to see how prices have changed over a given number of days.

  Parameters:

  • days — integer — Look-back window in days (default: 30, max: 365)
  • limit — integer — Maximum number of snapshot rows to return (default: 50, max: 500)
  • provider — string — Filter by provider (e.g. 'openai'). Omit for all providers.
  • model_name — string — Filter by model name (e.g. 'gpt-4o'). Omit for all models.

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_pricing_trends

  Find models whose prices changed the most over a given period. Returns models sorted by absolute percentage price change with a direction label ('increased', 'decreased', or 'unchanged'). Use this to answer questions like 'which models got cheaper this month?' or 'which providers raised prices recently?'

  Parameters:

  • days — integer — Look-back window in days (default: 30, max: 365)
  • limit — integer — Maximum models to return (default: 20, max: 100)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• register_price_alert

  Register a webhook URL to receive notifications when a model's price changes by more than a specified percentage. Optionally scope the alert to a specific provider or model. Returns the alert ID.

  Parameters:

  • url (required) — string — Webhook URL to POST when the alert fires
  • provider — string — Limit to a specific provider (e.g. 'openai'). Omit for all.
  • model_name — string — Limit to a specific model name. Omit for all models.
  • threshold_pct — number — Minimum absolute % price change to trigger the alert (default: 5.0)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• list_price_alerts

  List all registered price-change webhook alerts with their IDs and settings.

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• delete_price_alert

  Delete a registered price-change alert by its ID.

  Parameters:

  • alert_id (required) — integer — The ID of the alert to delete (from list_price_alerts or register_price_alert)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_pricing_export_url

  Generate a download URL for the pricing history export. Returns a URL the user can open in their browser to download snapshot history as CSV (for spreadsheets) or JSON (for code). Use this when the user asks to export, download, or save pricing data.

  Parameters:

  • days — integer — Look-back window in days (default: 30, max: 365)
  • limit — integer — Max rows to export (default: 10000, max: 100000)
  • format — string — Output format: 'csv' (default) or 'json'
  • provider — string — Filter by provider (optional)
  • model_name — string — Filter by model name (optional)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• list_conversations

  List stored chat conversation sessions with their IDs, last-updated timestamp, turn count, and a preview of the most recent user message. Use this when the user asks to see their past conversations or chat history.

  Parameters:

  • limit — integer — Maximum number of conversations to return (default: 20, max: 100)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• delete_conversation

  Delete a specific chat conversation by its ID. Use when the user asks to delete, clear, or remove a past conversation. Use list_conversations first to find the conversation ID.

  Parameters:

  • conversation_id (required) — string — The ID of the conversation to delete

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_ide_pricing

  Get subscription pricing for AI coding IDE tools: GitHub Copilot, Cursor, Windsurf, Claude.ai, JetBrains AI, Amazon Q Developer, and Tabnine. Returns monthly subscription cost, inline completion support, context window, and per-tool strengths. Use this when the user asks to compare IDE AI tools, subscription-based coding assistants, or wants to know which IDE AI fits their budget.

  Parameters:

  • provider — string — Filter by provider name (partial match, case-insensitive). Examples: 'GitHub Copilot', 'Cursor', 'Windsurf', 'JetBrains', 'Amazon Q', 'Tabnine'
  • inline_only — boolean — If true, only return tools that support real-time inline code completion
  • max_monthly — number — Only return plans at or below this monthly USD price (e.g. 20 for $20/mo)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• predict_cost

  Predict the real cost of an LLM call from your actual prompt text — no need to know token counts in advance. Counts tokens locally, infers the task type (or accepts one explicitly), estimates output length from empirical I/O profiles, and returns all models ranked cheapest-first with effective cost after prompt-cache savings. Also surfaces the best quality-per-dollar pick. Use this before making any LLM API call to avoid surprise bills.

  Parameters:

  • prompt (required) — string — The actual prompt text to cost-estimate (required)
  • top_n — integer — Number of models to return, ranked cheapest-first (default: 10, max: 50)
  • task_type — string — Task type for output-length estimation. Auto-inferred from prompt if omitted. Options: classification, extraction, summarization, qa, code_generation, translation, chat, content_generation, reasoning, function_calling, data_analysis, rewrite
  • require_vision — boolean — Only include models that support image/vision input (default: false)
  • cache_hit_ratio — number — Fraction of calls that hit the prompt cache (0–1). Set this if you have a large reusable system prompt. Supported providers: Anthropic (10% cache rate), OpenAI (50%), Google (25%). Default: 0
  • min_context_tokens — integer — Minimum context window size required in tokens (optional)
  • require_function_calling — boolean — Only include models that support function/tool calling (default: false)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• optimize_workload

  Optimise a multi-task LLM workload by assigning the cheapest qualifying model to each task instead of running everything through one model. Takes a list of workloads (task type + monthly volume + optional quality floor and capability requirements) and returns a per-task model allocation, the total monthly cost, and the saving versus the best single-model deployment. If a monthly budget is given, leftover headroom is spent on the most cost-efficient quality upgrades. Use when the user asks how to cut LLM spend across several task types, or whether tiered model routing is worth it.

  Parameters:

  • workloads (required) — array — One entry per task type in the workload mix
  • min_quality_score — number — Global quality floor (0-100) applied to every workload
  • monthly_budget_usd — number — Total monthly budget in USD. Leftover headroom is spent on the most cost-efficient quality upgrades.

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• recommend_model

  Recommend the optimal LLM model from a free-text description of your use case — no need to already know the routing engine's task_type vocabulary. Classifies the description into a task type, then runs it through the same routing engine behind /router/recommend (live pricing + benchmark quality, hard filters for cost/quality/context window, scored by quality-per-dollar). Returns the best match plus up to 3 alternatives. Read-only — use predict_cost first if you also want per-model cost estimates for an actual prompt, and optimize_workload if you have several distinct task types to route at once.

  Parameters:

  • description (required) — string — Free-text description of the use case, e.g. 'summarize long customer support tickets' or 'generate Python unit tests'
  • task_type — string — Optional explicit task type, bypassing auto-classification of the description. Options: classification, extraction, summarization, qa, code_generation, translation, chat, content_generation, reasoning, function_calling, data_analysis, rewrite
  • ide_context — string — IDE context hint — boosts IDE-native tools: 'copilot' | 'cursor' | 'windsurf' | 'claude_code' | 'jetbrains' | 'amazon_q'
  • avg_input_tokens — integer — Average input tokens per request (default: 500)
  • avg_output_tokens — integer — Average output tokens per request (default: 200)
  • min_quality_score — number — Minimum required quality score (0-100)
  • min_context_window — integer — Minimum required context window size (tokens)
  • monthly_budget_usd — number — Monthly spend ceiling in USD. Combined with estimated_monthly_requests, avg_input_tokens, and avg_output_tokens to hard-filter models that would exceed it.
  • prefer_low_latency — boolean — Prioritise lower-latency models (default: false)
  • preferred_provider — string — Preferred provider name (10% score boost applied)
  • max_cost_per_1m_tokens — number — Maximum acceptable average cost per 1M tokens (USD)
  • exclude_reasoning_models — boolean — Exclude slow thinking/reasoning models (default: false)
  • estimated_monthly_requests — integer — Estimated requests per month (default: 1000)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• check_price_drift

  Audit this server's own pricing data against an external reference registry. Prices here are hand-maintained and go stale silently; a price that disagrees with the registry by more than the threshold is automatically withheld from serving the moment it's detected, not just when this tool runs. This reports what's currently withheld and why. The price value itself is never overwritten by the registry. Use when the user asks whether the pricing data is accurate or up to date, or wants to audit a specific provider's rates.

  Parameters:

  • limit — integer — Maximum findings to return (default: 50, max: 200)
  • provider — string — Limit the audit to one provider (partial, case-insensitive)
  • threshold_pct — number — Minimum % difference to report (default: 5)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_data_quality

  Report an aggregate accuracy summary for the pricing catalogue: what percentage of models have a confirmed, fresh price, how many are currently withheld due to price drift, how many are newly discovered and not yet priced, and how many confirmed prices are over 90 days old. A single trust signal, distinct from check_price_drift which lists individual disputes. Use when the user asks how accurate or trustworthy the pricing data is, overall.

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_cache_stats

  Report process-lifetime hit/miss counts, total lookups, and hit rate for the ~5-minute recommendation cache, separately for the recommend_model tool and POST /router/recommend (never combined). Server-operations introspection for confirming the cache is reducing recomputation in production, not a pricing question. Counts reset on process restart; reading them never changes them.

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• record_usage

  Record actual token usage for a completed LLM call. Cost is computed server-side from current pricing, never trusted from the caller — this is the 'actual spend' counterpart to estimate_cost's hypothetical numbers. Optionally scope to an org_id for per-org spend tracking. Resubmitting the same request_id for the same org is a no-op. Use when the user wants to log or track real LLM spend, not just estimate it.

  Parameters:

  • model_name (required) — string — Name of the LLM model that was called
  • input_tokens (required) — integer — Number of input tokens actually used
  • output_tokens (required) — integer — Number of output tokens actually used
  • org_id — string — Optional organisation ID to attribute this usage to
  • request_id — string — Optional idempotency key — resubmitting the same value is a no-op
  • session_id — string — Optional session ID to tag this event with, so it can later be grouped and retrieved via analyze_session_usage — e.g. a bounded burst of calls like a coding-agent run or a chat session
  • occurred_at — number — Unix timestamp when the call happened (default: now)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• get_usage_summary

  Get a summary of actual recorded LLM spend: total cost, request count, and breakdowns by model and provider, from usage events previously logged via record_usage. Use when the user asks what they've actually spent, as opposed to a hypothetical cost estimate.

  Parameters:

  • days — integer — Look-back window in days (default: 30, max: 365)
  • org_id — string — Optional organisation ID filter

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• analyze_session_usage

  Analyze one specific session's actual recorded LLM usage (tagged via record_usage's session_id) on demand: total input/output tokens, total actual cost, and a per-model breakdown if more than one model was used. Also runs the session's own observed token pattern through the same routing engine behind recommend_model / POST /router/recommend to produce a grounded recommendation — or confirm the current choice is already optimal — with the estimated dollar savings or increase versus this session's actual cost. Returns has_data: false with no recommendation if the session_id has no recorded usage. Advisory only — never switches or reroutes anything.

  Parameters:

  • session_id (required) — string — The session_id previously used when calling record_usage
  • org_id — string — Optional organisation ID filter — scopes the lookup to usage events recorded under that org_id, same as get_usage_summary's org_id

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• import_session_usage

  Bulk-import a completed session's usage in a single call — the retroactive counterpart to record_usage, for when a session wasn't instrumented live. Use when a calling agent (e.g. VS Code Copilot Chat) already has a session's turn-by-turn history in its own local store and wants this server to report on it exactly as if it had been tracked live. Accepts a list of entries under one shared session_id; each entry resolves its input and output tokens independently, either from exact counts when known or by estimating from raw turn text using this server's existing token counter. Cost is always computed server-side from current pricing, never trusted from the caller. A malformed entry (no usable tokens/text on one side, or an unresolvable model_name) is reported as its own per-entry failure and does not fail the rest of the batch. Idempotent like record_usage — resubmitting the same entries (same request_id, or the same session_id/model/occurred_at/token-or-text when request_id is omitted) is a no-op. Recorded usage is then readable via analyze_session_usage and GET /usage/session/{session_id}.

  Parameters:

  • session_id (required) — string — Shared session ID to tag every entry in this batch with
  • entries (required) — array — Usage entries to import, one per turn/call in the session
  • org_id — string — Optional organisation ID to attribute this usage to

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• register_budget_alert

  Register a webhook URL to receive notifications when actual recorded spend (from record_usage) crosses a USD threshold within a rolling window. Optionally scope to a specific org_id (omit for total spend across all orgs). Fires at most once per period_days while spend stays over threshold. Returns the alert ID.

  Parameters:

  • url (required) — string — Webhook URL to POST when the alert fires
  • threshold_usd (required) — number — Spend level in USD that triggers the alert
  • org_id — string — Organisation to monitor. Omit for total spend across all orgs.
  • period_days — integer — Rolling look-back window for the spend check, in days (default: 30)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• list_budget_alerts

  List all registered budget-spend webhook alerts with their IDs and settings.

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• delete_budget_alert

  Delete a registered budget-spend alert by its ID.

  Parameters:

  • alert_id (required) — integer — The ID of the alert to delete (from list_budget_alerts or register_budget_alert)

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

• ask_agent

  Ask the LLM Pricing AI agent a question in natural language. Use for complex queries about pricing, model recommendations, cost optimisation advice, or anything that benefits from multi-step reasoning.

  Parameters:

  • message (required) — string — The question or task for the pricing agent
  • autonomous — boolean — Run in autonomous task mode (default: false)
  • conversation_id — string — Optional conversation ID to continue a previous session

  Hints: read-only not declared · destructive not declared · idempotent not declared · open-world not declared

Repository

Measured by MCPYP probing. Untested fields are omitted, never estimated. Listing is always free; MCPYP never proxies calls or stores user tokens.