Skip to main content
Trinity
Updated Nov 17, 2026main@87e21ff
FAQ/Sharing, Access & Monetization

Sharing, Access & Monetization

Sharing, roles, access requests, quotas, the Workspace Inbox, clients and sessions, paid agents. Short, grounded answers with links to the full documentation.

53 questions

How do I share an agent with a teammate?

Open the agent detail page, click the Settings tab and pick Peoplein the rail (visible to the owner; it is the former Access tab), enter your teammate's email, and click Add person. They can then sign in to the Workspace with a code emailed to them and use the agent there. Sharing does not create a platform account or add the email to the login whitelist — whether they also get the operator UI is their platform role, which an admin sets on their row or in Settings → Access. Their row shows Active if they already have a Trinity account, or Pending("No platform account — uses the Workspace"); each row also carries an off-by-default toggle that lets the agent message that person proactively. Click Remove on a row to revoke access at any time. See Agent Sharing.

What can my teammate do with an agent I've shared with them?

Shared operators get interact-level access: they can chat with the agent, run tasks, start and stop it, and open its files except credential files. They see only the executions they started, never the owner's or anyone else's, and they cannot modify the agent's configuration, credentials, schedules, or permissions, or delete or re-share it — those stay with the owner (and admins). The agent itself is held to their limits too: no shell, file tools limited to the paths they could open, no schedules or reminders set for them, and on a Gemini CLI or Codex agent their message is refused. See Agent Sharing.

Why can't my teammate see my past conversations with a shared agent?

That's by design. Shared users see only their own chat messages with an agent — each person's conversation history is private to them. Only admins can see all messages. If your teammate needs the content of a past conversation, share it with them directly. See Agent Sharing.

What's the difference between Settings › People and the Channels tab?

Settings › People(formerly the Access tab) manages the people the agent is shared with — your teammates, who use it in the Workspace, and in the Trinity UI too if their role is operator or above. The Channelstab manages external clients — people without a Trinity account who reach the agent through Slack, Telegram, WhatsApp, voice calls, public links, or the Workspace, identified by verified email. The Channels tab also holds the Restricted/Open access switch with pending access requests, the Public chat model and Additional instructionsthat apply only to outside audiences (public links, channels, paid chat — never your own chats or schedules), channel configuration, a read-only client roster, and the public links panel; file sharing lives under Settings › Storage. See Agent Sharing.

What user roles does Trinity have and what can each one do?

Trinity has four hierarchical roles: admin > creator > operator > user. Admins can create agents and manage every agent on the platform; creators can create agents and manage their own; operators cannot create agents but can work with agents assigned to them; a userworks in the Workspace only — the agents shared with them, their chats and the Inbox, with no Dashboard, agent pages, Operations or Settings. Higher roles inherit all permissions of lower ones. See Roles and Permissions.

Who can create agents?

Creating agents requires the creator role or above. If a teammate needs to create agents, an admin can promote them: go to Settings → Access → User Management, find the user in the table, and pick a new role from the dropdown — the change takes effect on their next request. You cannot change your own role, and admins cannot demote themselves. See Roles and Permissions.

Why am I getting an agent quota error when I create an agent?

Each role has a limit on how many agents it can own: by default creators get 10, operators 3, and users 1 (admins are always unlimited). When you're at your limit, agent creation is rejected with HTTP 429 and a message of the form Agent quota exceeded. You have 10/10 agents. Delete an agent to create a new one. (code QUOTA_EXCEEDED, with your current count and limit). Admins can raise the per-role limits under Settings → Agent Quotas (enter 0for unlimited). Redeploying an agent you already own doesn't count against your quota, and system agents are excluded from the count. See Agent Quotas.

What role does a new user get the first time they log in?

The role comes from their email whitelist entry: each whitelisted email carries a default role that is assigned on first login, and it falls back to the basic userrole if none was set. Adding someone through an agent's Settings › Peopleor approving their access request creates no account and no whitelist entry — they use the Workspace with an emailed code. A user account works in the Workspace only; an admin raises someone to operator or above via Settings → Access → User Management (or the role select on their row there) to give them the operator UI. See Roles and Permissions.

Why does everyone, me included, have to accept terms of use on first sign-in, and can I change them?

Every Trinity instance ships default terms of use, and each person signing in to the app or the Workspace accepts the current version once before they can continue — admins included. An admin edits the title and Markdown text under Settings → Access → Terms of use and clicks Save as a new version, which asks everyone to accept again on their next visit; Restore the default puts the shipped text back, and a switch turns the requirement off. The same panel lists who accepted which version and when. Scripts, the CLI and operator or agent API keys are never asked. See Setup → Terms of use.

Can people sign up for my Trinity instance on their own?

Not unless you explicitly allow it — public self-signup is off by default, and the email whitelist stays the real access gate. The unauthenticated access-request endpoint returns 403 and tells the person to ask an administrator to whitelist their email. An admin can opt in via the PUBLIC_ACCESS_REQUESTS_ENABLED environment variable or the matching system setting; when enabled, self-signups are auto-whitelisted with the basic user role. See Roles and Permissions.

How does someone request access to my agent, and what happens when I approve it?

When an agent is in Restrictedmode and an unknown user with a verified email messages it — from any channel — they see "Your access request is pending approval" and a request appears under the Restricted/Open switch in the agent's Channels tab. Click Approve to add their email to the share list, which admits them on every channel at once; click Denyto reject silently (the agent's existence is not confirmed). Each request names who asked as the channel saw them — name, @handle, the channel and their id on it, and their email — with up to three “seen before” badges showing where that same identity has asked, been approved or denied, or verified on your other agents. If the request came in over Telegram, Slack, or WhatsApp, Trinity automatically messages the requester on that same channel to confirm access; a delivery failure never rolls back the approval. See Access Control.

What's the difference between Restricted and Open access on an agent?

It's a single per-agent switch under "Who can chat with this agent?" on the Channels tab. Restricted(the default) means only the owner, admins, and explicitly approved emails can chat — everyone else generates a pending access request. Open means anyone with a verified email can chat immediately. Email verification is not a separate switch on that tab: both positions of the control require a verified email (it writes require_email: true either way), and turning verification off entirely (require_email: false) is possible only through the access-policy API (PUT /api/agents/{name}/access-policy). The fleet-wide default that newly created agents start with lives under Settings → Generaland applies at creation time only — existing agents are never rewritten. See Access Control.

What is the Workspace?

The Workspace at /workspaceis the signed-in chat app for the people you share agents with — and for you. It ships in every build (the old /portalpath redirects there), and it is the one surface where a conversation keeps the agent's working memory between turns, where every (you, agent) pair has a pinned Mainchat plus as many named chats as you like, and where several agents can work one thread together. Platform users open it from the nav — it opens in its own browser tab, and your platform session is your Workspace session; external clients sign in with a 6-digit email code, get no platform account, and see only the agents shared with their address. It opens on the Inbox — what needs you and what came back across your agents — and clicking an agent starts a new chat with it, with your earlier chats as tabs. See Workspace.

How do I bring a second agent into a conversation?

Type @and the other agent's name in an existing one-to-one chat. The Workspace opens a room containing both agents and posts your message there, leaving the original chat untouched. Inside a room, + Add agent(or mentioning an agent that isn't yet a participant) recruits it — only a person can do that, never another agent — and @ keeps working there while the / skill picker is not offered in rooms. An @namethat isn't one of your agents stays plain text. Multi-agent chat is part of the open-source platform — it used to be an enterprise capability, and against an older backend that lacks it the picker is single-select and mentions stay ordinary text. See Workspace.

What can a client see about an agent in the Workspace?

Clicking an agent opens the conversation, not a report about it. A band under the header of every chat with that agent shows its tasks in the last 7 days (a fixed window), the share completed, the share that succeeded first try, a Helpful / Not helpfultally, and a small activity chart. Everything else sits in the rail's Infotab: name and description with health and availability as separate facts, the agent's Role card when it has one, the client's own chats with unread counts, What it can do capability cards that pre-fill the composer, the reports addressed to them, What it remembers about you (with Undo for a run that changed it), and the Decisionsthey recorded with the agent. A client's band and Info tab count only runs they can see — their own turns, what those turns started, and the agent's shared scheduled runs — never other people's runs of the same agent. Beyond Info, a client gets Files (files exchanged in the chat) and Canvas (only canvases the agent published to its roster); the Work and Loopstabs, the model dropdown, and voice calls are platform-user-only, and a client never sees loop runs it couldn't open. Nothing configures — no schedules, skills, logs, costs, or model details — and the live updates a client receives are scoped to the agents on their roster. Because it reads from stored data, a stopped agent still renders. See Workspace.

What can I do with an agent's canvases from the Workspace rail?

The rail's Canvas tab shows one canvas at a time; pick another from the dropdown in the control row above it, where a pinned canvas carries 📌 and comes first, and once an agent has more than six a Search canvases…box filters by title or id. The header states two facts — Updated 2h ago · agent last ran 40m ago— so you can judge freshness yourself, and PDFprints the open canvas through your browser's own print dialog. The agent's owner (or an admin), signed in as a platform user, also gets Manage at the end of that row, which opens a list below it: each row shows its age, a pin toggle and Delete, and a checkbox per row feeds Delete selectedwith one confirmation naming the count; everyone else — including every external client — has a read-only panel. The canvas you have open travels with each message you send from that chat, so add a column to thisnames the right one; it is context, not permission. Sharing a canvas by link is done from the agent's Canvas tab on Agent Detail, not from the rail (next question). See Workspace.

How long does a client stay signed in to the Workspace?

A client session slides: it renews while the person uses the Workspace, ends after an idle window, and never outlives a hard cap — by default 7 idle days and 30 days in total. When it times out the sign-in form says Your session timed out, and signing in again returns them to where they were. Admins see the policy under Settings → Retention → Workspace sessions; changing the numbers requires an entitlement. Platform users have no separate Workspace session — their platform login is the Workspace login. See Workspace.

Why does the Workspace ask me to accept terms of use before I can see my agents?

The instance shows its terms of use once to every person who signs in, on any door — a client in the Workspace and a platform user alike — and again when the terms change. Tick the box and press Accept and continue, and the Workspace loads your agents, chats and asks. Decline and sign out signs you out and changes nothing else; sign in again when you are ready. An admin edits the text, or restores the shipped default, under Settings → Access. See Workspace.

Why does the Workspace say "Add an email to use the Workspace"?

The Workspace finds your agents — the ones you own and the ones shared with you — by email, and your Trinity account has none. This is usual for the built-in admin on a fresh install. Press Open Settings, add an address under Settings → General → Admin sign-in email, then Reload. Separately, if you type an email that belongs to a Trinity account into the Workspace's own sign-in form, you are sent to the Trinity sign-in page with it filled in, so you get your own agents rather than a client's view. See Workspace.

What's the difference between "Sign out" and "Sign out of Trinity" in the Workspace?

The button at the bottom of the Workspace sidebar signs you out. For a platform user it reads Sign out of Trinity, because the Workspace session isthe platform session — ending one ends the other. For an external client it is a plain Sign out. Identity never switches on its own: when a client's session expires on a browser that also holds an operator login, the sign-in form offers Continue as user@example.com as an explicit click rather than quietly carrying on as the operator. See Workspace.

Which model does a client's conversation run on, and can they choose it?

A client's turns run on the model you set as the Public chat model on the agent's Channels tab, falling back to the platform default; clients get no model control of their own. Platform users on a Claude-runtime agent do get a dropdown beside Send — Agent's default (…) plus three plain-language tiers, Most capable, Balanced — fast and smart, and Fastest— remembered per agent on their account. If the agent can't complete a turn on a chosen model, the reply says so and the choice reverts to the agent's default. See Workspace and Agent Sharing.

Can my agent deliver a report to one specific client, and what happens when they rate it?

Yes. The agent names the role the report is for with to — usually primary, the person it serves — and the platform resolves who that is; the older audience_email still works but is deprecated. When that person is on the agent's roster and is not its owner, the report becomes a deliverable: it appears in that person's Workspace under Info → Reports, and as a Delivered here card at the end of the chat whose turn produced it (the agent passes its execution_idand Trinity resolves the chat, so it can never post into a conversation it wasn't part of). Unaddressed reports never appear in the Workspace. Each deliverable carries Useful / Not what I needed; the second opens an optional comment box, one rating per person applies (clicking the other option corrects it), and a negative rating also raises a heads-up in the operator queue for the agent's owner. The agent learns tallies only — never the words and never who rated it; the agent's owner and admins read comments in full, and other people the agent is shared with never see the heads-up. See Agent Reports.

What is the Decisions section in the Workspace?

A record of what you and the agent approved, deferred or killed, and why. Each entry records the outcome, what was decided, the alternatives, the criterion that decided it, what would reverse it, and a review date. You record one with Record a decision, and the agent can record one with its record_decision tool. An active decision can be reconfirmed, corrected, reversed (with a reason) or closed; nothing is deleted, and a decision past its review date shows as expired. The active ones are read into every turn with that agent, so it reuses the criterion next time. See Workspace.

What does "What it may do unprompted" mean on an agent's Info tab?

It lists each kind of ask (the Ask class on your decisions) and whether the agent may act on it without asking you first. Every class starts on-request; it becomes unpromptedonly when at least three decisions from separate conversations apply one criterion with no reversals and you haven't rated the agent down in 30 days — there is no promote button. Even then, the instance's autonomy level must be at L2or above and the agent's autonomy switch on; each blocking reason is listed in plain words. You can Hold on-requestany class on your own seat, and only the agent's owner can release a hold. See Workspace.

Why does the role card say the agent "owns no objective yet"?

The objectives an agent owns come from its seat on record — the seat it holds itself, or the seat of the person it serves — not from the role named in its template. An agent with no seat on record owns nothing, so the card says so and still lists the objectives that name it as a supporting agent. The seat shown is the one the agent serves for you, with a line saying where it comes from. See Workspace → The role card and Dynamic Dashboards → Objectives.

Can my agent ask a client a question, and where does the client answer it?

Yes. An agent addresses an approval or question to one person it is shared with by setting addressed_to_emailon the queue item; an address that isn't on its roster turns the item into an ordinary operator ask rather than being refused. The client sees it in their Inbox; an ask raised during a chat turn also appears as a tile inside that chat, answerable there, while one raised outside any chat (by a scheduled run, say) appears in no chat. The sidebar counts open asks separately from unread replies (an orange ring on the agent's row), and while asks are waiting the agent's rail has an Askstab whose rows open each one in the Inbox. An ask to a role several people fill reaches all of them, and the first answer wins. Answering there records the decision exactly as an operator's would, including waking the agent when that setting is on, and the item stays visible to operators in the queue. The ask's context block is never shown to the client, and revoking the share hides the ask. See Approvals.

What is the Workspace Inbox?

The Workspace's landing page — signing in, or opening bare /workspace, takes you there. Action lists the asks waiting on you, most urgent first (soonest expiry, then priority, then oldest), and can be narrowed to one agent; Unread lists chats where something came back since you last read them; All lists every chat plus asks that ended in the last 7 days. Clicking a row opens it in a reading pane: an ask is answered in place, with where it came from shown below it, and a chat shows what arrived and is marked read. A pinned Inbox row at the top of the sidebar carries the two counts. When a colleague tags you in a conversation, that lands under Unreadtoo, and so does a room you are in when an agent posts there — one row per room, which opens its new messages in the pane. See Workspace → The Inbox.

How do I tag a colleague in a Workspace conversation?

Type @and start their name in the message field: once you type, the list offers people as well as agents — accounts on this instance that can already reach one of the chat's agents. Pick them and send. They get one Unreaditem in their Workspace Inbox (in the app only, no email), and under your message a mark shows whether it landed and whether they read it. A tag never wakes an agent and never grants access: the tagged person sees the message itself only if they can already see that conversation, otherwise they're told who can let them in. External clients can't tag people. See Workspace → Tagging a person.

Why does clicking an agent open an empty chat instead of my last conversation?

Most visits to an agent start new work, so opening an agent now starts a new chat with the cursor in the message field. Nothing is created until you send, so this leaves no empty chats behind, and if you left unsent text in a chat with that agent you land back on it. Your earlier chats are the tabs above the thread, and ⌥⇧↓ / ⌥⇧↑ (Alt+Shift+↓ / ↑) steps through them. Moving between agents with ⌥↓ / ⌥↑ returns you to the chat you last had open with each one, until you reload. See Workspace.

How do I close a chat tab, and can I get it back?

Click the ×on its tab. Closing archives the chat — nothing is deleted — and an Undotoast appears straight away. A closed chat leaves the tab strip and the sidebar's date groups and is listed under Archived at the bottom of the sidebar; opening it there reopens it. Mainhas no ×, and the open chat can't be closed while its reply is still coming. Pressing Reset on Main archives the old Main the same way. See Workspace → Chats as tabs and Main.

Does the Workspace have keyboard shortcuts?

Yes, and they work while you are typing in the message field. ⌘J / Ctrl+J starts a new chat, ⌥↓ / ⌥↑ moves between agents, ⌥⇧↓ / ⌥⇧↑ between chats with the current agent, ⌘. / Ctrl+. shows or hides the rail, ⌥. steps through the rail's tabs, ⌘/ / Ctrl+/ jumps to the sidebar search, and ⌥/ / Alt+/ opens the full list (also the Keyboard shortcuts button at the bottom of the sidebar). Esc closes the innermost thing first, then stops a running turn. The keys pause while a dialog is open, and during a voice call only ⌘J answers. See Workspace.

How do I reply to one specific message from the agent?

Click Reply to this messagein the action row under the agent's message (beside Copy message and the thumbs). A Replying to …chip appears above the message field, and your next message tells the agent which message you are answering; the chip's ×or Esc drops it. It works in 1:1 chats, and from a chat opened in the Inbox's reading pane. See Workspace.

How do I sign a client out of the Workspace, or block them?

On the agent's Channels tab, the Portal clients list shows everyone the agent is shared with. Log outends that person's live Workspace sessions everywhere — one sign-in covers all their agents — and they can sign straight back in; any owner of an agent shared with them can do it. Block is admin-only and keeps them out of the whole platform until an admin unblocks them, keeping their history. To remove someone from one agent only, unshare it. See Agent Sharing.

Are there limits on how much a client can send through the Workspace?

Yes, per client. Chat sends are limited per client per agent — 20 per minute and 300 per hour by default — and file sends per client at 20 per minute and 100 per hour, each answered with a 429 that names the reason (Too many messages to this agent., Too many uploads.); admins tune them with PORTAL_CHAT_BURST_LIMIT / PORTAL_CHAT_HOURLY_LIMIT and PORTAL_UPLOAD_BURST_LIMIT / PORTAL_UPLOAD_HOURLY_LIMITon the backend. A message can be up to 100,000 characters — the composer counts down over the last 500 and holds Sendpast the limit. Files are capped at 25 MB each and 20 per drop, turns on one chat are serialized (a second message waits for the one running), and sign-in codes carry brute-force protection. The agent's own execution timeout and parallel capacity apply on top. See Workspace and the single-server guide.

Can my own product sign a customer into the Workspace on their behalf?

Yes, with a Portal delegateMCP key. An admin mints it (human-only, via the scope choice on the create-key form); it reaches exactly one route — exchanging an end user's email for a Workspace session — so a trusted backend can act as that person and embed Workspace conversations in your own product, and every other path refuses it. The asserted email must already have an agent shared with it — the key cannot create access for anyone — and every exchange is audited. See Authentication.

How do I keep a growing fleet of agents organized?

Use tags and saved views. Tag agents from the agent detail page (or via API/MCP — POST /api/agents/{name}/tags/{tag} adds one, tag_agent is the MCP tool); tags show as colored badges on agent tiles and drive the shared Dashboard filters that apply across the Timeline, Grid, and List views and persist across navigation. On the Grid view, dept-<name> and reports-to-<agent>tags draw the org overlay — department zones and reporting lines — and those two families can only be set by people, never by agent-scoped keys. For filter combinations you use often, create a System View— a saved filter of tags plus other criteria that persists across sessions. See Tags and Organization.

Can I manage Trinity from my phone?

Yes. Trinity ships a mobile-optimized PWA at /m (for example http://your-domain.com/m) — install it via Add to Home Screen for a native-app feel. It has three tabs: Agents (list, start or stop, chat, toggle autonomy, view logs), Ops(answer the operator queue and acknowledge alerts — for an approval you tap an option, optionally add a note, then press an explicit Send, so nothing goes out on a single tap), and System (fleet health as Total / Running / Stopped / High Ctxcounts, plus the fleet-level actions — Emergency Stop, Fleet Restart, Pause Schedules, Resume Schedules — each confirmed in a bottom sheet). It's built for quick interactions like answering an agent's question or flipping autonomy on and off. See Mobile Admin.

Can I charge people to chat with my agent?

Yes, via per-request payments using the Nevermined x402 protocol. In the agent's payment settings, enter your Nevermined API key, Agent ID, and Plan ID, then enable payments — the agent gets a paid endpoint at POST /api/paid/{agent_name}/chat. Callers without payment credentials receive HTTP 402 with payment instructions, buy credits on the Nevermined checkout page, and retry with a payment-signature header; Trinity verifies the payment, deducts credits, and routes the message to the agent. If the agent is also exposed over A2A, the same price applies to outside callers on its A2A door, and its A2A card advertises the price. See Nevermined x402 Payments.

How do I set the price for my paid agent, and can I test payments without real money?

Pricing lives in your Nevermined plan (which defines credit pricing and allocation — one plan per agent); on the Trinity side you set how many credits each chat request burns (credits_per_request, default 1; set 0for a time-based plan, where the plan decides what a call costs). Crypto plans and card (fiat) plans both work — Trinity reads the plan's type from Nevermined and asks the buyer for the matching payment scheme, with nothing to configure. For testing, configure the agent with the sandbox environment (the default), which runs on the Base Sepolia testnet; switch to the liveenvironment for real payments on Base mainnet. Anyone can check an agent's payment requirements without authentication via GET /api/paid/{agent_name}/info. See Nevermined x402 Payments.

How can I see the payments my agent has received?

Payment history is available per agent via GET /api/nevermined/agents/{name}/payments, or through the get_nevermined_paymentsMCP tool. If a payment settles incorrectly, admins can list failed settlements and retry them manually — settlement failures are never retried automatically. See Nevermined x402 Payments.

My Nevermined plan is paid by card. Does that work?

Yes. Trinity reads the plan's type from Nevermined and advertises the matching x402 scheme in its 402 response — nvm:card-delegation for a card plan, nvm:erc4337for a crypto one — so the buyer's token is checked against the right requirements. The answer is cached for five minutes, and GET /api/paid/{agent}/infoshows which scheme a client will be asked for. If Nevermined can't be reached the first time, Trinity uses its last answer or falls back to the crypto scheme, and logs a warning. See Nevermined x402 Payments.

Why does a buyer's payment token get rejected for the wrong resource URL?

A 402 carries a resource.urlthat the token is minted and verified against, so it must be the address the buyer actually calls. Trinity uses your configured public URL when it matches the host of the request, otherwise the request's own host, upgraded to https only when your proxy sends X-Forwarded-Proto: https. Check that your TLS terminator forwards that header, and that the configured public URL is the hostname buyers use. See Nevermined x402 Payments.

Can other systems pay to call my agent over A2A?

Yes, when the agent is both exposed over A2A and has payments enabled. A caller with a Trinity key still runs free; a caller without one gets 402, pays with a Nevermined token sent in the A2A message, and gets a normal Task back with the payment receipt in its metadata. The agent's A2A card advertises the price so a client can pay on its first request. A price alone does not open the door — A2A exposure must be on. See A2A Protocol.