Atolio works where you do.

Atolio ships with prebuilt connectors for the most widely used enterprise platforms and provides a custom connector framework for proprietary or less common systems. Out-of-the-box connectors include Google Workspace (Drive, Gmail, Calendar, Meet), Microsoft 365 (SharePoint, Teams, OneDrive, Outlook), Confluence, Jira, Slack, Salesforce, GitHub, Zendesk, Notion, Box, Dropbox, and others. The full catalog is available on Atolio’s connectors page.
‍
For sources not covered by a prebuilt connector, Atolio's custom connector framework lets your engineering team build and deploy connectors against Atolio's ingestion API. Custom connectors follow the same indexing and permission model as prebuilt connectors, so content from proprietary systems participates fully in unified search.
‍
Custom connector development requires engineering resources on your side. Atolio's implementation team provides documentation, reference implementations, and technical support during the build.

Yes. Atolio can index content from external-facing and third-party data sources, not only internal collaboration and productivity tools. Supported third-party source types include public web content indexed via web crawl connectors, external documentation portals, partner-shared repositories, and external SaaS platforms that expose APIs for content retrieval, such as vendor knowledge bases or regulatory databases.
‍
For third-party sources requiring authentication, Atolio's connector framework supports OAuth, API key, and service account authentication models. Content from authenticated external sources is indexed with access controls you configure, making it visible only to the internal user populations you specify.
‍
Connecting third-party content requires a configuration review to confirm that indexing is consistent with those sources' terms of service and with your organization's data governance policies. Atolio's implementation team advises on connector configuration for complex third-party scenarios.

Atolio's Salesforce connector indexes across the major Salesforce object types, not just record names, making Salesforce knowledge searchable alongside content from all other connected sources. The connector indexes Accounts, Contacts, Leads, Opportunities, Cases, Tasks, and custom objects. For each record type, Atolio extracts structured fields (name, stage, amount, close date) alongside unstructured content (notes, email threads, activity history, attachments). This makes queries like 'what did we promise Acme in the last QBR' answerable from natural language.
‍
Salesforce permission enforcement is applied at the object and field level, matching the sharing rules and field-level security configured in your Salesforce org. Users only see records and fields they are authorized to access in Salesforce itself.
‍
The connector supports standard Salesforce orgs and Salesforce for financial services and healthcare clouds. Custom objects and custom fields are supported through connector configuration during implementation.

Yes. Atolio connectors can be built using the SDK to retrieve data from internal data sources. For on-premises environments, Atolio deploys a connector agent within your network perimeter. The agent communicates with on-premises source systems over your internal network without requiring those systems to be internet-accessible, then forwards extracted content to the Atolio indexing pipeline in your private cloud.
‍
For legacy databases or mainframe content stores, the custom connector framework lets your team build an extraction pipeline that delivers content to Atolio's ingestion API.
‍
On-premises connectivity requires network configuration to allow the connector agent to reach both the source system and the Atolio indexing endpoint. Atolio's implementation team documents and validates required network paths during deployment planning.

Atolio's connectors implement rate limit management for every supported source system, ensuring Atolio's indexing activity does not disrupt the availability of source systems for end users. Handling strategies include exponential backoff for API calls that hit rate limit thresholds, time-windowed request budgets that spread indexing activity across hours to avoid burst limits, priority queuing that deprioritizes bulk historical crawl traffic relative to ongoing delta sync, and per-connector configuration options to set conservative API consumption caps.
‍
Initial full-history indexing generates the highest API load. Atolio's implementation team reviews your source system API quotas during deployment planning and configures the initial crawl schedule accordingly. For source systems with restrictive quotas, a phased crawl over several days is standard.
‍
Ongoing delta sync is designed to operate within normal API quota headroom for enterprise-scale deployments. Atolio can provide an API consumption estimate for any specific source system during the scoping process.

Atolio keeps the index current through continuous delta sync: each connector monitors its source system for changes and propagates updates to the Atolio index on a rolling basis, without full re-crawls. Delta sync works by tracking modification timestamps, webhook events, or change logs depending on what each source system exposes. When content is created, modified, or deleted, the relevant connector detects the change and updates the index accordingly. Deletion events and access revocations are handled the same way: affected content is removed or suppressed within the next sync cycle.
‍
Last-sync timestamps and error states for all connectors are visible to administrators in the Atolio admin dashboard.

Atolio indexes both structured and unstructured data and ranks them together in a single unified results list: a Salesforce record, a Confluence page, and a Slack message are scored by the same relevance model. For structured data sources, Atolio extracts field values and maps them into a normalized document schema. A Salesforce Opportunity is indexed with its structured fields (name, stage, amount, close date) alongside unstructured content (notes, email threads, attachments), making both findable through natural-language queries.
‍
For unstructured data, Atolio applies text extraction, chunking, and embedding generation. PDF, Word, PowerPoint, and HTML content types are processed into searchable text, with document structure preserved where possible to improve retrieval quality.
‍
Both content types are represented in keyword-searchable and vector form, regardless of source format, so hybrid search applies equally to a database record and a long-form document.