Plans and limits
What each SourceLace plan includes, how AI use is counted, how you pay, and what happens at each limit. Your organization's plan is set by your SourceLace contact; admins see it on the Connections, People, AI model, Usage and Billing pages.
Prices are not listed here because they can differ by contract: your organization's prices, including any discount you agreed, are on the Billing page.
Features your plan does not include are never hidden: the app shows them with the plan they are available on, such as "available on the Business plan".
The plans
The standard plans at launch. SourceLace can change them over time, so your organization's own plan, as it applies to you, is the one shown in the app (the Billing page names it, and the Limits page shows its limits).
| Starter | Team | Business | Enterprise | |
|---|---|---|---|---|
| Seats (people who can sign in) | 1 to 5 | 5 to 50 | 25 to 500 | 75 and up |
| How people use SourceLace | Their own AI app (Claude, ChatGPT, Cursor) | Their own AI app and the SourceLace app | Their own AI app and the SourceLace app | Their own AI app and the SourceLace app |
| AI questions included per seat a month, pooled | None | 100 | 150 | 200 |
| Standard connectors included | 3 systems | Unlimited | Unlimited | Unlimited |
| Changes to records (preview and confirm) | Yes | Yes | Yes | Yes |
| Tool calls per person a month (fair use) | 5,000 | 10,000 | 25,000 | 50,000 |
| Access by group, single sign-on, audit log viewer and export | No | No | Yes | Yes |
| SCIM provisioning | No | No | Yes | Yes |
| Audit history kept | 90 days | 1 year | 3 years | 7 years |
| Longest chat retention | No app chat | 90 days | 1 year | 10 years |
| AI agents on at once | None | 5, drafts and summaries only | 50 | Unlimited |
These are the standard plans at launch. Your SourceLace contact can change any of these for your organization, for example for a pilot. SAML single sign-on is part of Enterprise; OpenID Connect single sign-on is part of Business and Enterprise. Older audit entries are deleted once they are older than your plan's audit history.
Data protection rules (hide fields and mask values before they reach the AI) are part of every plan. Runners, which run chosen agents inside your own network, are offered on the Enterprise plan only, on request, and are turned on by your SourceLace contact. See Runners, onboarding and Premium support.
Other limits, such as how long an agent may run or how long people stay signed in, are on the Limits page; admins can make some of them stricter.
Your pilot
Before you sign a contract, your SourceLace contact can start a pilot: a set time in which your organization tries SourceLace on your own systems. No card is needed, and nothing is billed. A pilot has the plan's features with a few limits, set by SourceLace: by default it lasts 30 days, allows 5 people and 5 sources, includes AI in the SourceLace app up to $50 in all (then chat in the app stops; your own AI apps keep working), and runs 5 agent. Changes to your records are allowed in a pilot, but each one still needs the usual approval first. Your SourceLace contact can set different limits for your organization. Your admins see how many days are left, and SourceLace emails them before the pilot ends.
If the pilot ends before the contract is signed, SourceLace does not close your organization: sign-in is paused and tells people the pilot period ended and to ask SourceLace to extend it. Nothing is deleted. Ask your SourceLace contact to extend the pilot or to continue with a contract: you can sign in again at once, with everything as you left it. When the contract is signed, your plan's own limits replace the pilot limits.
Seats
A seat is one named person on the People page, however they joined (invited, by email domain, through single sign-on or by SCIM). People your identity provider deactivated take no seat. The People page shows how many seats are used.
- Starter and Team: at the plan's maximum, nobody new can be invited or join by email domain until someone is removed or more seats are added. Admins see a notice at 90% of the maximum.
- Business and Enterprise: you can go over the maximum; the extra seats are counted and billed at the quarterly true-up.
Connectors
Connectors are grouped in three classes. Standard connectors are included in your plan, up to a number of systems. Every Business and Enterprise connector is a separate charge for each system: no plan includes them, and each shows as its own line on your contract, so you can see what each costs.
| Class | Connectors |
|---|---|
| Standard | Gmail and Google Calendar (google_workspace), Google Drive, Google Sheets, Microsoft 365 mail and calendar, SharePoint and OneDrive, Amazon S3, Azure Data Lake and Blob Storage, Google Cloud Storage, PostgreSQL, MySQL, SQL Server, MongoDB, Jira, Zendesk, HubSpot, custom connectors, any REST or SOAP API, outside MCP servers |
| Business | Salesforce (with Data 360), Marketing Cloud (Pilot), Dynamics 365, Dynamics 365 Finance and Supply Chain, ServiceNow, NetSuite, Gong, Gainsight, ChurnZero, Flexport, Snowflake, BigQuery, Databricks, Oracle Database, Trino and Starburst, Microsoft Fabric and Azure Synapse, Amazon Redshift, Amazon Athena, lake tables (Iceberg, Delta, Parquet) |
| Enterprise | SAP S/4HANA, SAP SuccessFactors, SAP ECC on-premise, Oracle Fusion Cloud, Workday, SAP Datasphere, health records (FHIR), healthcare claims (Stedi, Availity, Optum) |
One more system of an included connector is a monthly add-on, and each system of a Business or Enterprise connector is a monthly charge for that connector; the Billing page shows the prices for your organization.
A system is one source: two Salesforce orgs are two Business systems. Your plan includes a number of systems of the Standard connectors, and your contract lists each system of a Business or Enterprise connector by name ("SAP S/4HANA x 1"). On Connections, connectors you can add are listed as usual; the others are marked locked, with the reason: what a system costs, or "Ask your SourceLace contact". A pilot or demo organization is never billed, so it can connect these without them being on a contract. SourceLace checks this when a source is added, however it is added (the app or an AI app), so the limit cannot be passed by accident. Your contact can also switch a connector off for your organization, for example while it is in testing; its sources then stop working until it is switched on again.
Runners, onboarding and Premium support
A runner runs chosen AI agents inside your own network. Runners are offered on the Enterprise plan only, on request, with these terms:
- A runner fee a year, per runner, on top of your subscription. Each runner is a Production or a Test runner; a Test runner costs less. Your contract says how many of each your organization has, and your admins can add up to that many on the Runners page.
- A one-time onboarding fee, for installation help and a security review.
- Premium support, a yearly line: faster response times, named contacts and a scheduled upgrade plan. Runners need it.
- A minimum contract term of one year once runners are on your contract. Until the contract's end date, the plan, the runners on the contract and Premium support stay as agreed. At the end date your runner licenses end, unless your contract renews automatically (License end dates). If your contract renews automatically (the Billing page shows it), the next term starts by itself and prices do not rise unless you agree to it; otherwise nothing renews by itself and your SourceLace contact extends the contract with you.
Prices are not published: ask your SourceLace contact. The Billing page shows the runner and Premium support lines with your prices, and the onboarding fee once it is charged. If your organization stops meeting these terms, its runners get no new work until it does again; see When runners get no new work.
AI use: credits and questions
AI use in the SourceLace app is counted in credits, each worth a small fixed amount. Each question uses credits in proportion to the AI work it needed, so a short question uses fewer credits than a long one, and a question on a larger model uses more than the same question on a smaller one. The app shows credits as questions: a typical question's worth of credits counts as one. The AI use section (on the AI model page) shows how many credits make one question.
- Included. Each plan includes a pool of questions per seat a month, shared by the whole organization: heavy users borrow from light ones. Billed seats are the most of the seats bought, the plan's minimum and the people on the People page. The pool renews on the 1st of each month.
- Prepaid questions. Packs of questions, bought ahead, are used after the included questions and before any extra use. The Billing page shows the size and price of a pack.
- Extra use. After that, each question has a price (shown under AI use on the AI model page, and on the Billing page), up to a monthly maximum your admin sets there. By default the maximum is a share of the included questions; 0 turns extra use off. A maximum above the included amount needs a card on file (Team) or your SourceLace contact's approval (Business and Enterprise). Extra use is a separate line, "Extra AI questions", billed after the month ends.
- At the maximum, chat in the SourceLace app pauses until the 1st of next month, or until an admin raises the maximum (that applies at once). SourceLace keeps working in your own AI apps: they do their own thinking and use no credits.
- Your own AI key. With your organization's own key from Anthropic, OpenAI, Google Gemini, Mistral, Azure OpenAI or Amazon Bedrock (AI model), your provider bills you directly and questions use no credits.
Admins see a banner in the app at 80% and at 100% of the included questions, and at 80% of the maximum extra use. At 80% and at 100% of the included credits, each admin also gets one email that month for each. The Usage page shows this month's credits, and Billing shows each month's credits, prepaid use and extra questions. Until online billing is set up for your organization, these are estimates.
AI spend: where credits go, budgets and advice
The AI spend tab of the Usage page (admins) shows where your organization's AI credits go, a month at a time. Choose this month or any of the 12 before it.
- The month against your plan. The credits used so far out of those your plan includes, with a mark at 80% (or the share your admin chose under Limits), and a chart of the credits used each day. While the organization is at or past that mark, the tab shows a note, and each admin gets one email at 80% and one at 100% each month.
- Where the credits went. Five tabs, each searchable, sortable and shown 25 rows at a time: Where (app chat, agents, agent tests, Search everything, the setup assistant, other), Group (your access groups), Agent (every agent, not only shared ones; click one to see its runs), Person and Model. A person's spend counts toward each of their groups, so the groups can add up to more than the organization's total. Export CSV downloads the tab you are on for the chosen month.
- Budgets. An admin can set a monthly budget in credits for a group or an agent. At 80% (or your chosen share) the admins, and an agent's owner, get one email and see a warning on the tab. At 100% the budget either warns only or stops AI use until the 1st, as the admin chose: a stopped group's people cannot start chat questions or other AI work, and a stopped agent's runs and tests do not start. Raising or removing a budget applies at once. Agents' owners see their own agent's budget on its page.
- Ways to spend less. When an agent's tests scored the same or better on a cheaper model your organization allows, the tab and the agent's page say so, for example "Renewal nudges scored the same on claude-sonnet-5-5 in its tests: about 48% fewer credits". The saving is an estimate from the agent's recent runs and current prices. Nothing changes by itself: Compare again queues the same test run on both models, and the agent's owner decides whether to change its model.
An agent's page (its Runs tab) shows its credits this month with the same daily chart, its budget, and any advice. Its owner, co-owners and admins see it.
Engine rows
When a question asks across systems, or needs more rows than a normal answer holds, SourceLace fetches rows from your systems to work out the answer. Those are engine rows. Each plan includes a number of engine rows a month for the whole organization (100 million unless your contract says otherwise). They use no credits and cost nothing extra: the rows your plan includes are the only measure. An answer a system works out entirely by itself uses none.
- The Usage page shows "Engine rows this month: X of Y", how many are left and when they renew (the 1st of each month).
- At 80%, each admin gets one email that month, and admins see a note in the app, the Usage page included, with how many rows are left and when they renew.
- When none are left, questions that need to fetch more rows are refused until the 1st, with a message saying so. Questions that fetch nothing keep working. To get more, ask SourceLace to raise your plan's engine rows.
Before a very large question runs, SourceLace shows its plan with about how many rows it would fetch and how many are left, and asks first. See Asking across systems.
Billing
Admins open Billing in the menu. It shows:
- How you pay. Starter and Team pay by card: click Pay by card to go to Stripe's secure payment page (Team can choose yearly or monthly seats). Business and Enterprise are invoiced, payable within 30 days; invoices arrive by email from Stripe. Manage card opens Stripe's page to change the card or billing email.
- What you pay. Each line of your contract (platform fee, seats, connector add-ons, and on a runner contract the Production and Test runners and Premium support) at list price and at your price, with any discount or agreed price (for example, seats at a set price each month), and the total a year. Extra AI questions and prepaid packs show their price each. A one-time charge, such as the runner onboarding fee, is shown below the total, as owed until it is invoiced.
- Prepaid questions. Card customers can buy packs here; they are added as soon as the payment goes through. Invoiced customers ask their SourceLace contact, who sends an invoice.
- Invoices. Your latest invoices from Stripe (up to 24, newest first), with each one's status and links to view it online or download the PDF. The links sit under the invoice number.
- AI use by month. Questions, credits, prepaid use and the "Extra AI questions" line for each month, at list price and at your price.
The invoices and the month-by-month list are tables like the Agents list: drag a column heading's edge to resize it, and use Columns to move or hide a column. The amounts are the same as before.
Extra AI questions are billed after each month ends, on a separate invoice line "Extra AI questions". Seats over a Business or Enterprise plan's maximum are added up and billed after each quarter.
If a payment fails or an invoice is past due, admins see a banner linking to the Billing page. Nothing is switched off: SourceLace always gives 30 days' written notice before any change to your access.
Your SourceLace contact may not have set up online billing for your organization; the page then shows estimates and says so.
Your contract
The Billing page shows Your contract for admins, read-only: the term, the contract dates, when it renews, whether it renews by itself and the notice period, and, if your contract has them, the date until which prices are fixed, the most the price may rise at renewal, the most seats you may drop at renewal, and whether unused AI credits roll over. To change any of it, ask your SourceLace contact. SourceLace emails your admins before a pilot ends, before the contract renews, and (when it renews by itself) before the notice period ends.
Unused credits. If your contract lets unused AI credits roll over, credits you did not use in a month are added to the pool of the following months, up to the limit and for the number of months in your contract. Credits that are about to expire are used first. The Usage page shows how many credits rolled in.
Fair use
Each plan has a fair-use number of tool calls per person a month (every lookup, from the app or an AI app). Nothing stops when someone passes it; admins see a notice, and repeated heavy use may be discussed at renewal.
Locked features
- Starter: the SourceLace app's chat is available on Team and up. Everything else works through your own AI app, as described in Getting started.
- Starter and Team: access by group, single sign-on and the Audit log page are available on Business and up. Every call is still recorded in your audit trail and counted on the Usage page. If you had groups or single sign-on before, they keep working; you can turn access by group off and remove SSO at any time.
Retention
Your admin chooses how long saved chats are kept, up to your plan's longest setting (see the table). A longer setting is refused with the plan that allows it. If your plan changes to one with a shorter limit, chats are kept for at most the new limit. The audit trail is kept for your plan's audit history; older entries are deleted automatically, and the entries kept still verify.