Notion Custom Agents vs Town: Shared Pages or Personal Routines?
Choose between Notion Custom Agents and Town by where context lives, how routines execute and which approval settings govern shared work.
Founder, Task Machine
Notion Custom Agents and Town give you a concrete context decision: should a recurring job read your team's pages and databases, or operate through a personal assistant with an editable Wiki? A weekly company readout can fit either model. The distinction becomes important when you need the routine to repeat an exact sequence.
Town documents code routines with saved step results that survive an approval pause. Notion's agent guidance recommends letting a Custom Agent determine how to reach an instructed outcome, with delegation available. Choose the execution method your report actually needs before comparing how each assistant remembers your preferences.
Pages as the shared starting point
Notion Custom Agents automate work using pages, databases and connected apps. Its agent guide describes shared agents that run from schedules or events, retain memories and split parts of a job among sub-agents. Your team can keep the input documents and the agent's work in the same workspace.
A team that already updates a task database can use that context for routing and reporting. Personal Notion Agent provides a separate on-demand experience, including turning a conversation into a launch plan with tasks and owners. Do not assume that every personal-agent funding option also applies to a shared background agent.
Notion also documents objectives with owners, measurable key results and calculated progress. That supports a workspace where people maintain the goal records. Its public pages do not describe a named agent lead responsible for reviewing and coordinating each goal, project or team.
An assistant with context you can edit
Town gives each person a Townie, with work retained in threads containing messages, responses and tool activity. Its context and memory guide describes persistent global and routine-specific memories plus an editable Wiki. You can inspect, add or delete memories and correct the context directly.
Its team model shares integrations, skills and background routines while preserving personal assistants. Town's public pages do not describe a shared work-item record that a person or agent can be assigned to own. A thread is useful for steering a conversation, but verify whether it gives your team the handoff record it needs.
For business reporting, the Company Goals Tracker reads numbers from connected trackers, compares targets and produces a weekly private readout with owners and blockers. That is a specific documented routine. It should not be mistaken for an assigned lead role across all projects or goals.
Instructions versus saved execution steps
Notion's Custom Agent guidance recommends clear outcomes, defining done and manual testing. Its docs do not establish a fixed step-and-branch process for Custom Agent runs. Workers add custom code in beta on Notion's servers, so keep their release status visible in an evaluation.
Town's code routine documentation describes exact steps for filtering, grouping data and calling AI helpers. Saved results allow a run to resume after an approval pause without repeating completed steps. For a report that first filters records and then groups them, inspect those intermediate results during evaluation.
The comparison concerns execution structure, not a promise of correct conclusions. Notion recommends reviewing run outputs and logs. Town validates action arguments and allows review before selected actions, but its public pages do not describe general evaluation of delivered results against defined requirements or rubrics. A valid tool call can still produce a report that a person must check.
Approval defaults and the plan you buy
Both products let you restrict tools. Notion's connection guide permits automatic tool execution or an approval request. Its security documentation also describes approval for unexpected external links.
Town's modes and approvals guide places requests in chat and Notifications, with per-tool Ask each time settings. New routines start in Auto, and team routines always use Auto mode. You can still override individual tools, so inspect the actual sending or publishing tool before sharing a sensitive routine.
Notion's pricing page shows Business at €19.50 per member per month billed yearly. Purchased Notion credits cost $10 per 1,000 monthly or $13 per 1,000 annual credits. Its subscription funding guide permits Business and Enterprise members to connect an eligible Plus, Pro, Business or Go subscription for selected GPT models in personal Notion Agent, but that connection does not work with Custom Agents or Auto.
Town's plans guide lists Starter at $15 monthly with 1,250 credits and two custom daily routines. Pro is $49 per user per month billed annually with 5,000 monthly credits and optional overage at $0.044 per credit. Starter's daily schedules do not include the live event triggers available on Pro and higher, which matters if a new email must start work immediately.
Notion Custom Agents and Town, point by point
These answers cover the same questions for both vendors, with a source and check date for each. The explanations distinguish personal context, shared work and the limits attached to particular plans.
Both columns answer the same questions from each company's public pages. Yes means the company publishes the capability the question asks about, Partly means something narrower, No means it says it does not offer it, and Not published means its public pages do not describe it. Every answer links to its source and shows when we last checked it. The oldest check on this page was November 13, 2026.
| What you compare | Notion Custom Agents | Town |
|---|---|---|
| Goals |
Notion Custom Agents
Partly
Notion documents objectives with owners, measurable key results and progress tracking. Its example database calculates progress from current, initial and target values, rather than requiring an agent to own the goal. |
Town
Yes
Company Goals Tracker gives your Townie a weekly review of company goals. It reads numbers from connected trackers, compares them with targets and reports movement, owners and blockers in a private readout. |
| Leads |
Notion Custom Agents
Partly
Managers can oversee projects and team workloads, and Custom Agents can delegate parts of a job. Its docs do not establish a named lead role responsible for reviewing and coordinating a goal, project or team. |
Town
Not published
Its docs describe no assigned lead role that reviews and coordinates a project, goal or team. Team Owner and Admin roles govern access and administration, while background routines can triage tickets. |
| Tasks |
Notion Custom Agents
Yes
Notion tasks have assignees, dates and status. The separate Claude agents beta accepts assigned tasks and keeps sessions and activity on the page or task where the agent works. |
Town
Partly
Each piece of work has a thread with messages, assistant responses and tool activity. Its docs describe no shared work-item record that either a person or an agent can be assigned to own. |
| Recurring work |
Notion Custom Agents
Yes
Custom Agents run on recurring schedules, Notion changes and Slack events, including messages, mentions and reactions. You can combine trigger types and filter when an agent runs. |
Town
Yes
Routines start on timezone-aware schedules, incoming email, calendar events or authenticated HTTP webhooks. Starter permits daily schedules, while Pro and higher plans support live event triggers. |
| Workflows |
Notion Custom Agents
Partly
Custom Agents follow instructions and hand work to sub-agents, and Workers add custom code in beta. Its docs do not establish a fixed step-and-branch process for agent runs, and recommend letting the agent determine how to reach the outcome. |
Town
Yes
Code routines follow exact steps each time, including filtering, grouping data and calling AI helpers. Completed steps are saved within a run so an approval pause can resume without repeating them. |
| Autonomy |
Notion Custom Agents
Yes
Custom Agent connections let you set individual tools to run automatically or always ask. Users receive confirmation prompts before actions, and unexpected external links generate a notification for approval. |
Town
Yes
Thread settings, personal routine modes and per-tool Ask each time controls govern approvals, collected in chat and Notifications. New routines start in Auto, and team routines always run in Auto mode. |
| Learned autonomy |
Notion Custom Agents
Not published
Its docs do not describe proposals to change an agent's permissions based on your approval history. Memories retain context, while approval settings control when the agent asks. |
Town
Not published
Its docs describe no proposal to change permissions based on your approval history. Progressive trust means you change routine modes or tool settings yourself, and saved memories do not change approvals. |
| Memory |
Notion Custom Agents
Yes
Custom Agents can retain viewable memories across chats and update them from corrections. Personal Notion Agent also keeps a Memories database where you can edit or delete saved context. |
Town
Yes
Your Townie uses persistent global and routine-specific memories in later work. You can inspect, add and delete memories, and edit its Wiki directly to correct the context it uses. |
| Planning |
Notion Custom Agents
Yes
Personal Notion Agent can turn a chat request into a launch plan with tasks and owners. Custom Agent creation also lets you review and refine draft instructions, triggers and access before saving. |
Town
Yes
You develop a routine in conversation, producing instructions, triggers, tools and a visible configuration preview that you can refine. Routines go live immediately, so this is not a required plan-approval gate before execution. |
| Intake |
Notion Custom Agents
Yes
Task routing agents capture incoming work and route it to a team. Slack messages, mentions and reactions can start a Custom Agent, while connected mail can trigger work from incoming email. |
Town
Yes
Email to your assistant and HTTP webhooks route requests into routines. A paid team's background routine can categorize incoming tickets and route them to the right owner. |
| Connectors |
Notion Custom Agents
Yes
Custom Agents support a catalog of hosted Model Context Protocol connections and custom servers, with scheduled and event-triggered runs. The Agent APIs also let integrations start chats, submit actions and read session history. |
Town
Yes
Town connects custom Streamable HTTP MCP servers and API-key services such as Cursor. Email events and authenticated HTTP webhooks can start routines using connected tools. |
| Models |
Notion Custom Agents
Yes
Auto selects models per task and can switch models during demanding steps of a longer job. It is the default for personal Notion Agent and Custom Agents, subject to workspace model restrictions. |
Town
Yes
Town says it selects the appropriate AI model for each job rather than requiring you to choose a provider. |
| Cloud or your own computer |
Notion Custom Agents
Partly
Custom Agents run inside Notion, and Workers run custom code on Notion's servers. Its docs do not describe running these agents on your own computer or infrastructure. |
Town
Partly
Town is a cloud service with a Mac client that reads iMessages and transcribes speech on-device. Its public pages describe no customer-hosted agent runtime, and local file access is listed as coming next. |
| Your own subscription |
Notion Custom Agents
Partly
Business and Enterprise members can connect a ChatGPT Plus, Pro, Business or Go plan to personal Notion Agent for selected GPT models. This does not work with Custom Agents or Auto. |
Town
Partly
The Cursor integration uses your Cursor API key to start cloud coding agents. Its docs describe no way to cover your Townie's own model usage with your AI subscription or provider key. |
| Pricing |
Notion Custom Agents
Yes
Business is €19.50 per member per month billed yearly on the fetched pricing page. Agent usage draws credits, with purchased Notion credits at $10 per 1,000 monthly or $13 per 1,000 annual credits. |
Town
Yes
Starter costs $15 monthly with 1,250 credits and two custom daily routines. Pro is $49 per user per month billed annually with 5,000 monthly credits and optional overage at $0.044 per credit. Free has a rolling 30-message allowance. |
| Correctness |
Notion Custom Agents
Partly
Notion recommends defining done in instructions, manually testing runs and reviewing their outputs and logs. Its docs do not establish automatic checks of delivered results against defined requirements or evaluation criteria. |
Town
Partly
Town validates action arguments and offers human review before selected actions. Its docs describe no general evaluation of delivered results against defined requirements or rubrics. |
For software work
| What you compare | Notion Custom Agents | Town |
|---|---|---|
| Code review rules |
Notion Custom Agents
Partly
The separate Claude agents beta accepts instructions for coding style, branch names and pull request descriptions. Its docs describe teammates reviewing changes, not an automatic code review against repository rules. |
Town
Partly
The assistant can inspect commit changes and review pull-request diffs through GitHub tools. Its docs describe no code review against configured repository rules. |
| Adversarial review |
Notion Custom Agents
Not published
Its docs do not describe an independent second review of risky code changes. The Claude task-board guide recommends keeping your normal teammate review of the diff and merge. |
Town
Not published
Its docs describe no independent second review of risky code changes. The Cursor integration starts coding agents and checks their progress rather than documenting a separate review gate. |
| Risk rating |
Notion Custom Agents
Not published
Its docs do not describe a risk rating for a code change before merge. The Claude task-board guide describes flagging risky scope in a product requirements document and leaving merge approval to a teammate. |
Town
Not published
Its docs describe no pre-merge risk rating. GitHub context cards track pull requests and deployment status and let you merge or enable auto-merge. |
| Pull requests |
Notion Custom Agents
Partly
The separate Claude agents beta opens GitHub pull requests linked to Notion tasks and accepts direction through task comments. Its docs do not establish handling review feedback from GitHub pull requests. |
Town
Partly
Town can commit file changes, create pull requests, read review comments and track status in a GitHub context card. Its docs do not establish an agent revision loop that addresses review feedback. |
A fact we cannot find or cannot check is left out, and prices can change after the date shown. Task Machine publishes this comparison and is an interested party. Notion Custom Agents and Town are named only to identify them, and there is no relationship between the companies. Has something changed? Tell us.
Who each one suits
Pick Notion Custom Agents if your team keeps the input context in pages and databases and wants background routing or reports there. Its public documentation recommends defining done in instructions, manual tests and output review, so budget time for that work.
Pick Town if you want editable personal context and code routines with exact steps that can resume after approvals. Confirm the paid trigger tier and the tool settings you need, since team routines use Auto mode even though individual tools can still ask.
Where Task Machine fits
Task Machine gives each job an owned Task with its plan, decisions and run history. Its planning documentation describes reviewing the plan and acceptance criteria before an agent starts, and Workflows can return failed step results through verifier feedback. Those boundaries matter when a shared review record is the requirement behind your search.
Task Machine is not the right fit if the job is already satisfied by a personal assistant's Wiki or by automation attached to existing Notion pages. Read Task Machine vs Notion Custom Agents and Task Machine vs Town for those work-record differences. Evaluate one weekly readout by pausing it before its final send and checking which completed steps and decisions you can inspect.