Hosted MCP server for LLM cost estimation, model comparison, and budget-aware routing.
Connect TokenOracle to Claude, Cursor or any other MCP client and it stops being a tab you switch to. Hosted MCP server for LLM cost estimation, model comparison, and budget-aware routing. The tokenoracle mcp server is what makes that connection.
Token Oracle is a Model Context Protocol (MCP) server that estimates, compares, and controls LLM API costs before agents spend tokens. It exposes nine tools, four read-only Resources, and a cost_analysis_workflow Prompt template. It uses a proprietary pricing algorithm without a backing LLM to ensure deterministic budget workflows.
The toolset is worth reading before you wire it up, because it tells you what the integration is really for:
estimate_cost — Estimates the USD cost of a single LLM API call before execution. Input: task_description, prompt_text, task_type, or explicit token_count. Outputestimate_cost_batch — Prices up to 100 LLM tasks in a single call. Returns per-task breakdown, total_cost_usd, and cheapest_model_for_all. Use before starting anycompare_models — Ranks LLM pricing across all supported providers for a given task. Returns models sorted by cost with speed_tier and quality_tier. Supports filteringbudget_check — Checks whether a planned task fits within a monthly budget. Returns can_proceed (boolean), remaining_budget_usd, budget_consumed_pct, andfind_cheapest_for_budget — Inverse of budget_check. Given a budget_usd cap and task, returns the best model/quality combination within budget plus all alternatives ranked bytoken-oracle — //meta — Machine-readable server capability document (version, model_count, pricing metadata)cost_analysis_workflow — Guided three-step prompt template: estimate_cost → compare_models → budget_check. Arguments: task_description, monthly_budget_usd (optional)Free — 1000 API calls/day, all hosted tools via the remote endpoint, no credit cardTOKEN_ORACLE_API_KEY — optional hosted API key; overrides any stored credentialTOKEN_ORACLE_BASE_URL — override for the remote endpoint; defaults to https://mcp.guffeyholdings.com/TokenOracleTOKEN_ORACLE_SUBJECT — optional end-user subject forwarded as X-Token-Oracle-Subjectlist_request_activity — get_usage_summaryBecause this one is hosted, setup is mostly authentication — you point your client at the endpoint and approve access. Nothing runs on your machine, so there is no runtime to keep patched.
Configuration is passed through the environment: TOKEN_ORACLE_API_KEY, TOKEN_ORACLE_BASE_URL. Treat anything key-shaped as a real credential — scope it to the minimum the server needs, and rotate it if it ever lands in a shared config.
This sits in the AI and media services group, where several servers overlap in what they claim to do but differ sharply once you actually set them up. TokenOracle's toolset — estimate_cost, estimate_cost_batch, compare_models and 9 more — is a fair guide to whether it matches your workflow. It is maintained by com.guffeyholdings; worth a glance at recent repository activity before you build anything load-bearing on it.
This entry was verified against TokenOracle's own documentation before publication; SyncDev keeps the directory reviewed rather than auto-generated.
| Tool | What it does |
|---|---|
| estimate_cost | Estimates the USD cost of a single LLM API call before execution. Input: task_description, prompt_text, task_type, or explicit token_count. Output: cost_usd, recommended_model, confidence, will_fit_context, pricing_updat |
| estimate_cost_batch | Prices up to 100 LLM tasks in a single call. Returns per-task breakdown, total_cost_usd, and cheapest_model_for_all. Use before starting any multi-step pipeline. |
| compare_models | Ranks LLM pricing across all supported providers for a given task. Returns models sorted by cost with speed_tier and quality_tier. Supports filtering by min_quality, max_cost_usd, and provider. Input: task_type, token_co |
| budget_check | Checks whether a planned task fits within a monthly budget. Returns can_proceed (boolean), remaining_budget_usd, budget_consumed_pct, and cheaper_alternatives with savings_pct. Input: monthly_budget_usd, current_spend_us |
| find_cheapest_for_budget | Inverse of budget_check. Given a budget_usd cap and task, returns the best model/quality combination within budget plus all alternatives ranked by quality then cost. |
| token-oracle | //meta — Machine-readable server capability document (version, model_count, pricing metadata) |
| cost_analysis_workflow | Guided three-step prompt template: estimate_cost → compare_models → budget_check. Arguments: task_description, monthly_budget_usd (optional), quality_threshold (optional). |
| Free | 1000 API calls/day, all hosted tools via the remote endpoint, no credit card |
| TOKEN_ORACLE_API_KEY | optional hosted API key; overrides any stored credential |
| TOKEN_ORACLE_BASE_URL | override for the remote endpoint; defaults to https://mcp.guffeyholdings.com/TokenOracle |
| TOKEN_ORACLE_SUBJECT | optional end-user subject forwarded as X-Token-Oracle-Subject |
| list_request_activity | get_usage_summary |
{
"mcpServers": {
"token-oracle": {
"type": "streamable-http",
"url": "https://mcp.guffeyholdings.com/TokenOracle",
"headers": {
"X-API-Key": "${TOKEN_ORACLE_API_KEY}"
}
}
}
}Configuration as documented by the project. Restart the client after saving.
| Variable | Description | Required |
|---|---|---|
| TOKEN_ORACLE_API_KEY | Credential the server authenticates with. | Yes |
| TOKEN_ORACLE_BASE_URL | Credential the server authenticates with. | Yes |
Build a programmable telecommunications stack for connecting telephony services with the Internet via a cloud-based utility.
Search built for AI, not humans — semantic web search that returns model-ready content, plus code context.
Answers, not links — delegate questions to Perplexity's search-grounded models and get cited responses back.
Give your assistant a voice — text-to-speech, voice cloning and audio tools from the ElevenLabs API.
Give your assistant a real code sandbox — isolated cloud VMs for actually running the code it writes.
The ML hub in your context window — search models, datasets, papers and run Spaces from the official server.