Claude
- Open Add custom connector ↗
- Name it Elis Token Saver and paste the URL
- Click Add, then sign in to your account
https://api.tryelisai.com/api/mcp/openai
Plan tasks. Track AI usage.
Elis Token Saver helps you plan work into smaller tasks and see how your AI host uses models and tools to complete it. Use it for writing, research preparation, coding plans, and other work that benefits from a clear sequence of steps.
Start with a goal. Elis returns a quote with an estimated saving before the host commits to the plan. When the host proceeds, Elis supplies the available tasks, their dependencies, and guidance about suitable model sizes and tools. The host carries out the work with its own available capabilities, reports results, and advances the same plan until it finishes. When a task needs clarification or cannot run, the host should explain the limitation and report it accurately.
Elis is a planning and measurement connection. It does not itself execute your work on Elis-hosted or community miners, and a model recommendation does not guarantee that your AI can switch models or launch agents in your current session. Available host tools, permissions, and model controls determine what can actually run. Savings estimates and usage-window estimates are distinct from measured results; actual outcomes depend on the task, host support, usage reporting, and your account settings.
Connect an Elis account through OAuth. Access and billing use the personal or organization scope selected during authorization. Use an existing Elis account with sufficient Token Saver entitlements. If an operation is unavailable to your account, Elis explains the limit and links to informational documentation. This connection does not administer refunds, change account passwords, or grant access to another customer's plans.
Elis receives the goal, task instructions, reported results, and usage information supplied by the host. Successful plans may be retained privately for reuse according to the published privacy policy. Ask to exclude a plan from reuse or community sharing when needed. Community sharing requires separate opt-in. You can also preview host setup instructions without applying file changes. Read the linked documentation, privacy policy, and terms for the current service conditions and contact support for account help.
Use Elis to write a three-question bicycle shop FAQ, then show every usage and fee report line exactly as returned.
https://api.tryelisai.com/api/mcp/openaihttps://api.tryelisai.com/api/mcp/openaiapi.tryelisai.comThe AI reads these server instructions before using this connector, every time. From the ChatGPT store listing.
Elis Token Saver supports user-requested task planning and usage reporting. Use it only when the user asks for Elis planning or measurement, or explicitly opts in to using Elis for the current work. Do not invoke it for unrelated requests or require it before answering other questions.
For an authorized workflow, returns a free quote. If you know how this host is billed, also pass host_billing: subscription (a ChatGPT plan seat or a session signed in with a ChatGPT account), api (pay-per-token API usage), or unknown. It only changes which report-card line leads and gates nothing; never read credential files to decide, and omit host_billing when unsure. If decomposition_request is present, submit the minimal task graph with . A quote_only request must stop without committing. commits the plan and reserves its quoted fee: proceed only within the user's execution and spend authorization. Report actual task outcomes to that same plan_id and current graph_revision, then call . Do not fabricate token measurements, savings, or feedback. Follow explicit no_store and no_community choices.
At the end of every authorized workflow, print the response's report_card.lines to the user verbatim, in order, immediately after the plan_id: do no math of your own and do not reword or drop a line. Distinguish estimates from measured results. If the response has no report_card (an older server), print: Usage extension: not provided by this Elis server
Within an authorized Elis workflow, if the user asks for a depth or model-size breakdown, pass expansion with max_depth: 0-3 and/or target_model_size: nano|small|frontier. The server applies speed-mode limits and returns the resolved expansion. Otherwise omit expansion to use the saved default.
When a task includes an eligible refinement_offer, its executor may split it instead of running it: report status refined and pass replan_request with reason, refine_task_id, and proposed_tasks (at least two) in the same call, only at a wave boundary. If refinement is refused, execute the original task.
When reported-savings charging is on, settlement uses the reported per-model saving; absent or unpriceable usage falls back to the reserved planning estimate, flagged for admin review. Otherwise settlement uses the reserved estimate. Explain the actual returned fee and settlement status.
Not found; the server needs sign-in so there may be skills that our scan couldn't access
Plan tasks. Track AI usage. Elis Token Saver helps you plan work into smaller tasks and see how your AI host uses models and tools to complete it.
Open its listing in the app's connector directory and connect. Sign in to your account in the browser window that opens. It uses OAuth, so the AI never sees your password. Step-by-step setup for each app is on this page.
The Elis Token Saver MCP server URL is https://api.tryelisai.com/api/mcp/openai. Paste it as a custom connector in Claude, ChatGPT, Gemini or Grok, or add it to Claude Code, Cursor or any MCP client. Setup steps.
Yes. Elis Token Saver is published or run by Noah Salas, so it is the official MCP server. See other official connectors.
It is listed in 1 store: ChatGPT. Apps where it is not listed can usually still add it as a custom connector.
It is ranked 15,399 of 15,719 in the ChatGPT directory. A rank is the store's own order, not a usage count. How we track ranks.
It has 5 tools: 1 read data, 2 change data and 2 can delete or overwrite data. See every tool and its inputs.
Yes. 2 tools can change data and 2 tools can delete or overwrite it. Most AI apps ask before they run these. The store marks these as destructive:
None that we found. Not found; the server needs sign-in so there may be skills that our scan couldn't access.
Here are examples from its store listing. Type one into your AI after you connect.
No. Its answers come back as text in the chat. Connectors with interactive UI.
This page updates every day. To ask about Elis Token Saver from Claude, ChatGPT or a coding agent, connect the Agent Ready MCP server: it tracks ranks, tools and new listings across the AI plugin stores.