Brisk Teaching helps educators create and adapt instructional materials directly from web pages, PDFs, videos, and Google Docs. Generate lesson plans, presentations, quizzes, rubrics, reading-level adaptations, and personalized student feedback aligned with grade levels and learning objectives.
Hermes is Nous Research's command-line agent. MCP servers are declared in its YAML config and reloaded without a restart.
This guide connects Brisk Teaching MCP to Hermes through MewCP, so Hermes can use Brisk Teaching MCP's 9 tools, like create_boost_activity, get_boost_activity and finalize_boost_activity, without your Brisk Teaching MCP credentials ever reaching the app.
Brisk Teaching MCP signs in through managed OAuth. MewCP stores the token encrypted, refreshes it before it expires and injects it only at call time. Your agent never holds it, so it can never leak it.
More apps
Manage courses, assignments, rosters and student activity within Google Classroom. Create learning workflows, track submissions and organize educational content across classrooms.
AI's impact on real-world sectors, as a structured record for agents: source-linked headlines placed in one of 18 sectors (healthcare, law, retail, government, defense and more), with outlet, country and time, since June 2026
The Synthesize Bio MCP Server enables AI agents to compare gene expression between sample groups, analyze bulk and single-cell RNA sequencing data, and run long-running bioinformatics workflows that return completed analysis reports.
Search Microsoft Learn documentation and training content through MCP, helping AI agents find relevant technical guidance, tutorials, modules, and learning resources.
Other clients
Brisk Teaching MCP works the same way in every client below. Pick yours for its own guide.
Free to start.
Scope access to the exact Brisk Teaching MCP tools and accounts you allow, work and personal side by side. Anything outside that scope is refused at call time, even when the model asks for it.
Each call is logged: what ran, on which account, and whether it worked. Failures come back structured, so the agent knows whether to retry instead of guessing.
Every app sits behind one secure MCP URL, and there is no server for you to deploy or keep running. Add an integration in the dashboard and it is live on the next call, with no config to change and no new credentials to wire up. Tools load only when a task needs them, so your agent stays fast and focused however many apps you connect.
Edit config.yaml in your Hermes config directory (~/.hermes).
Put this under the top-level "mcp_servers:" key.
Run /reload-mcp in Hermes, or start a fresh session. Sign in when the browser window opens.
Start a new chat and ask: "List the tools you can use from MewCP." You should see your connected apps.
Create a draft Brisk Boost activity for the connected teacher. A Boost activity is an AI-guided student experience (tutor, quiz, debate, exit ticket, etc.) generated from the teacher's prompt. This creates an editable draft and returns links to edit it in the Boost editor and to share it with students. The activity stays a draft until the teacher finalizes it; students only reach it once its link is shared. UI-capable hosts (e.g. Claude) render an interactive card — a loading skeleton paints immediately (the tool's UI resource bakes it in), then the finished card replaces it when this returns. Every client also gets the plain-text summary so headless agents and `client.py` can read the result.
Fetch the current state of a Boost activity the connected teacher owns. Read-only. Returns the activity's raw fields (draft/shared flag, links, objectives) — no card. The Boost card calls this on mount to refresh its live state; the model can call it to re-check an activity without re-creating or finalizing it.
Finalize a draft Boost activity — the card's "Share with students" step. Flips the activity's setup to complete so it appears as a finished activity (idempotent). This is a UI/lifecycle flag, not a student-access gate — students reach the activity only once its link is shared.
Generate a set of Brisk Boost activity ideas (interactive, AI-powered learning experiences) for a topic — e.g. tutor chats, debates, brainstorms, exit tickets. Read-only: generation creates nothing. The ideas come back as a carousel the teacher can browse; each card can be edited as a draft or created and shared with students in one click (those buttons call create_boost_activity / finalize_boost_activity). Use this when a teacher wants suggestions for a topic rather than a single named activity. UI-capable hosts (e.g. Claude) render the interactive carousel — a loading skeleton paints immediately, then the ideas replace it. Every client also gets the plain-text summary.
Set up a Brisk Batch Feedback assignment for grading a class set of student work. This creates the assignment shell only — the teacher adds the student submissions and runs the AI feedback in the Brisk web app (the card's link), not in this chat. Use this when a teacher asks about grading student work or giving feedback on a set of submissions. Feedback style and grade level are required (the rubric styles also need the teacher's rubric) — ask if unclear rather than guessing. Also required: which class to file the assignment under. Unless the teacher already named one, call get_teacher_context first and ask with their class names as options; pass the choice as class_section_name or class_section_id, or no_class_section=true if they pick none. UI-capable hosts (e.g. Claude) render an interactive card — a loading skeleton paints immediately, then the finished card replaces it. Every client also gets the plain-text summary.
The teacher's classes in Brisk (names, ids, grade levels, subjects) — never any student data. Call this at the start of a teaching task to personalize defaults like grade level, and before setting up an assignment so you can ask which class to file it under — offer the class names as options, plus 'no class'. Read-only. Returns the raw class context (no card); every client also gets the plain-text summary.
Look up education standards (Common Core, NGSS, AP, state standards, etc.). Read-only. Use this to find the standards a teacher wants to align an activity to, then pass the chosen `standard_id`s into the create tools. It is adaptive: call it with what you know and it returns either the available filter options for the next step (when the request is too broad) or the matching standards. A typical flow from "a 5-question quiz on multiplication for 3rd grade, Common Core": call with `standard_set="USA - National"`, `standard_set_subcategory="Common Core"`, `subject="Math"`, `grade="3"`, then narrow the returned list to the multiplication standards. Returns no card — structured data plus a plain-text summary the model can read directly.
Create a teaching resource (lesson plan, quiz, rubric, and 15 more types) for the connected teacher. One tool fronts Brisk's whole generation library: pick `resource_type` and Brisk writes the material from the teacher's prompt into a Google Doc (or a Google Form, for quizzes). The artifact is created up front and content streams into it, so this returns immediately with a link; the model can poll `get_teaching_resource` to read the streamed state. Ask the teacher for any required field (grade level; quiz/spiral question count; syllabus weeks; unit days; Form-vs-Doc for a quiz) rather than guessing. Only pass standards the teacher confirmed. UI-capable hosts (e.g. Claude) render an interactive preview card; every client also gets the plain-text summary.
Read a teaching resource's current streamed state from Google. Read-only. Forms parse into structured questions; Docs return the document text streamed so far. The preview card calls this on mount and polls it while content streams in; the model can also call it to check progress.
