# Guru Pricing Breakdown: Packages, AI Credits, and Knowledge Base Costs

Guru now sells quote-based packages rather than published per-seat plans, and it meters AI work through a shared pool of credits for each workspace. Evaluating the real cost requires accounting for package scope, credit consumption, and the ongoing work of keeping verified cards current alongside connected sources. Comparing Guru with intelligent workspace alternatives reveals how direct document indexing changes knowledge management economics.

Source: https://fast.io/resources/guru-pricing/
Author: [Derek Labian](https://fast.io/authors/derek-labian/)
Last reviewed: 2026-11-09

## How Guru Packages Plans, AI Answers, and Knowledge Base Tiers

Knowledge bases fail when maintenance costs more than the answers are worth. Teams evaluating Guru often discover that the total cost of ownership extends beyond the package price, driven by AI credit consumption, the scope of the package, and the ongoing labor of card-level verification.

Guru pricing is quote-based. As of June 2026, Guru's pricing page presents packages built from three elements (the Platform, the Expertise, and the Infrastructure) with no published prices, and AI work draws on a shared pool of credits for each workspace. Understanding what you will actually pay requires looking at how Guru packages its software, how credits are consumed, and what ongoing administration demands from your team.

### Public Pricing Shifts and Enterprise Packaging

Guru historically maintained public self-serve tiers, but the company has shifted its commercial presentation to custom proposals. On the official [Guru pricing page](https://www.getguru.com/pricing), Guru organizes packages into three core elements: the Platform, the Expertise, and the Infrastructure.

The Platform layer covers automated knowledge quality and maintenance, Knowledge Agents for cited, permission-aware answers, AI search, more than 100 integrations with MCP delivery, verification workflows, and deep research. The Expertise layer bundles a team of solution engineers who assist with knowledge architecture design, agent configuration and tuning, and rollout playbooks. The Infrastructure layer covers enterprise governance, role-based access controls, encryption, and audit logs.

Because current packages are quote-only, prospective buyers must engage directly with sales to receive a proposal. Guru's older published prices are still useful context, but they describe plans Guru no longer lists, so treat them as history rather than as a current quote.

### Guru's Published Pricing History

Archived copies of Guru's pricing page show how its published pricing changed before the move to packages:

| Period (Archived Pricing Page) | Published Plan | Annual Billing | Monthly Billing | Seat Minimum |
| --- | --- | --- | --- | --- |
| April 2024 and May 2025 snapshots | All-in-one | $15 per user per month | $18 per user per month | None listed |
| January 2026 snapshot | Self-serve, AI credits included | $25 per seat per month | $30 per seat per month | 10 seats |
| June 2026 snapshot and current page | Packages (Platform, Expertise, Infrastructure) | Not published | Not published | Not published |

The 2024 and 2025 All-in-one plan bundled AI search across connected sources, an intranet, and a wiki. The January 2026 self-serve plan listed AI credits with usage limits, Knowledge Agents, AI search and chat, a verified knowledge base, and a customizable intranet. None of these prices appear on Guru's current pricing page, so use them only as background when you request a quote.

## The Anatomy of Guru AI Pricing: Credit Pools, Packages, and Seats

The package price is only the entry ticket. In modern AI-enabled knowledge management, how a vendor meters automated answers dictates your ongoing operational budget.

Guru meters AI work in credits. Every workspace receives a shared pool of credits, and credits are not assigned to individual users. When a team member runs a search or asks a question, the platform searches verified cards and connected sources and draws from that pool: one search or one generated answer consumes one credit, and asking a Knowledge Agent to create a card, where it searches, drafts, and publishes, consumes three.

### How the Credit Pool Behaves

Because the pool is shared, heavy users draw down the same balance as everyone else. Support agents, sales engineers, and operations staff who search all day, and Knowledge Agents that create or update cards, consume credits faster than occasional readers.

Guru notifies admins as the workspace approaches its allocation, and more credits can be added if usage grows. During scoping, ask Guru how large the pool in your package is and what additional credits cost.

### Human Seats Versus AI Consumption

Guru's earlier published plans were priced per user or per seat, and its current packages are scoped to your organization's scale. Ask how the package counts users: whether everyone who reads cards, runs searches, or receives answers in Slack or Microsoft Teams needs a license, and how contractors and partners are handled.

Per-user licensing creates financial tension when the audience for company knowledge grows beyond employees. Exposing the knowledge base to external contractors, partners, or temporary project staff can add license cost. For teams developing autonomous workflows, evaluating dedicated [storage for AI agents](/storage-for-agents/) provides a more flexible alternative.

### Enterprise Governance in the Package

Guru's current packages include an Infrastructure layer for enterprise governance, which its pricing page lists as including:

- SSO, SCIM, and role-based access
- Data Loss Prevention (DLP) masking
- Data encryption at rest and in transit
- Permission-aware answers with citations and lineage
- Centralized audit logs across human and AI consumers

Because these controls and the Expertise layer's solution engineers are scoped into each package, ask which elements your quote includes. A package sized for enterprise governance and expert support is a larger commitment than a departmental tool.

## Cards, Connected Sources, and the Cost of Verification

When evaluating knowledge management platforms, finance and IT leaders typically compare subscription prices. The larger long-term cost in Guru is usually verification. Guru's model has two layers: cards, which people (and now Knowledge Agents) write and verify inside Guru, and connected sources such as Google Drive, SharePoint, Slack, and Zendesk, which Guru searches alongside the cards. The cost question is how much of your knowledge has to live in verified cards to be trusted, and who keeps those cards current.

Guru was originally designed around atomic, bite-sized units of information called cards. While modular cards are easy to share, real business documentation does not naturally exist in short snippets.

### Where Card Maintenance Still Costs Time

Modern businesses run on long-form documents. Technical specifications, legal contracts, vendor service agreements, employee handbooks, and operational runbooks routinely span dozens of pages. Financial models and inventory matrices live in extensive spreadsheets.

Guru can search these files through connected sources, so they do not have to be rewritten to be found. Turning them into verified answers is a different matter: content that needs a verifier, a verification cadence, and a verified badge has to become a card, whether a person writes it or a Knowledge Agent drafts it.

That card layer introduces its own upkeep:

1. Card Scope: Deciding which policies from a long employee handbook become standalone verified cards, and which stay as a connected source document, is an editorial job.
2. Loss of Context: When long-form documents are split into cards, the broader context of surrounding sections can be lost, leading to incomplete answers.
3. Duplicate Upkeep: When a source document in Google Drive or OneDrive changes, any card that restates it needs an update and a fresh verification, whether a person or a Knowledge Agent drafts the edit.

### The Verification Treadmill

To ensure information does not become stale, Guru implements a verification workflow. Every card has an assigned verifier (an individual subject matter expert or a designated team). A card can be set to come up for review every 30, 60, or 90 days, every 6 months, or every year, or set to not expire for evergreen content, and Knowledge Agents can verify or unverify cards that fit their criteria during daily runs.

When a card with a review interval reaches its expiration date, its status indicator turns from a green verified checkmark to an unverified warning. The assigned verifier receives automated notifications prompting them to review the content and reaffirm its accuracy.

In theory, this cadence maintains high knowledge quality. In practice, it creates administrative exhaustion. As an organization creates thousands of cards across multiple departments, subject matter experts face mounting verification queues. When engineers, product managers, and HR leads become overwhelmed by review notifications, verification backlogs grow. Once a significant percentage of cards display unverified badges, team trust in the knowledge base drops, reducing employee engagement and defeating the purpose of the platform.

## Calculating Total Cost of Ownership Across Knowledge Base Deployments

To understand the true cost of adopting a card-based knowledge base, organizations must calculate total cost of ownership. This calculation encompasses package fees, credit consumption, and the internal labor required for content authoring and verification maintenance.

### Cost Drivers Across Typical Team Deployments

Guru does not publish package prices, so the useful comparison is which cost drivers grow as headcount and content volume expand:

| Team Deployment Tier | Software Cost Drivers | Administrative Labor | Questions for the Quote |
| --- | --- | --- | --- |
| Small Team | Package price for a small team | Initial collections, first cards, and verifier assignments, then routine reviews | Does Guru offer a package at this size, and how large is the credit pool? |
| Mid-Sized Department | Package price plus a credit pool sized for support and sales search volume | Ongoing verification queue management across collections | What do additional credits cost, and how are readers licensed? |
| Enterprise Rollout | Package with Infrastructure governance and Expertise services | Multi-department verification queues and agent configuration | Which Expertise services and governance features are in scope? |

For a small team, the package is only part of the investment. Setting up collections, writing or approving the first cards, and assigning verifiers takes concentrated time, and recurring card reviews demand regular manager attention.

In a mid-sized department, heavy search use by support and sales teams draws down the shared credit pool faster, so the size of the pool and the price of additional credits matter. Managing collections, verifier assignments, and stale card queues across multiple projects can turn into a part-time role.

At the enterprise level, packages bundle the platform, solution engineering, and enterprise governance, and the internal labor of maintaining verified cards across many departments sits on top of the package price.

### Knowledge Management Architecture Comparison

The table below contrasts the commercial and structural characteristics of card-based wikis, traditional enterprise search, and intelligent workspaces:

| Capability | Card-Based Wiki (Guru) | Traditional Enterprise Search | Intelligent Workspaces (Fast.io) |
| --- | --- | --- | --- |
| Pricing Model | Quote-based packages with a shared AI credit pool per workspace | Per-seat or custom infrastructure index | Flat workspace plans with included credits |
| Document Ingestion | Cards written by people or Knowledge Agents, plus connected sources such as Google Drive, SharePoint, Slack, and Zendesk | API connectors crawling external drives | Direct file storage and scheduled Cloud Sync |
| File Types Supported | Cards, uploaded files, and content from connected sources | Indexes remote files in place | Native PDFs, spreadsheets, docs, and notes |
| Search Retrieval | Semantic search over cards and connectors | Enterprise keyword and semantic search | Auto-indexing RAG with specific file citations |
| AI Agent Integration | MCP server, Knowledge Agents, and Slack and Microsoft Teams integrations | Dedicated search portal or intranet | Remote MCP server for Claude, Cursor, and custom agents |
| Maintenance Overhead | Card verification on each card's review interval | Ongoing connector and permission mapping | Low direct file storage with version history |

The difference is where trusted answers live. Guru can search connected files, but its verification model applies to cards. Platforms that store and index the documents themselves cite the source file directly, rather than a card that restates it, which preserves the full context of the original business materials.

## Intelligent Workspaces as a Company Brain Alternative

Instead of maintaining a separate layer of verified cards alongside the source documents, modern teams are adopting intelligent workspaces as a [company brain](/product/company-brain/). In this architecture, documents live directly inside the workspace where both people and autonomous AI agents collaborate.

### Direct File Indexing Without Snippet Authoring

On Fast.io, knowledge management begins with your actual files. When an organization creates a workspace with Intelligence enabled, incoming files are automatically indexed for keyword and semantic search, structured extraction, summaries, and conversational chat.

There is no need to copy sections out of a 60-page PDF manual or summarize financial spreadsheets into individual cards. The system ingests raw PDFs, Word documents, Excel spreadsheets, presentations, and technical notes directly. When a team member or AI assistant queries the workspace, answers cite the exact file name, page number, and text snippet from which the information was drawn. This citation trail allows readers to inspect the source material immediately, providing verification without recurring manual card reviews.

### Ingestion via Direct Upload and Scheduled Cloud Sync

Files arrive in an intelligent workspace through two primary methods: direct file upload or Cloud Sync. Cloud Sync connects Google Drive, Dropbox, Box and OneDrive folders to a workspace, one-way or two-way, on a schedule or on demand. When SharePoint libraries need to be indexed, they connect through the OneDrive connector.

This synchronization ensures that when an engineering team updates an architectural spec in Google Drive or an HR manager uploads an updated benefits policy in OneDrive, the workspace ingests the new file version according to its configured schedule. Because synchronization can run automatically on a defined cadence, internal teams never need to maintain duplicate card repositories or manually paste revisions across systems.

### Remote MCP Server for Humans and AI Agents

Knowledge is most useful where work happens. Guru delivers answers through integrations and its own MCP server; Fast.io takes the same protocol approach for the files themselves. The Fast.io remote MCP server exposes workspace content to external agents using Streamable HTTP:

- Claude applications (claude.ai, Claude Desktop, Claude Cowork) connect at `https://mcp.fast.io/mcp/tools`
- ChatGPT and Codex custom server configurations connect at `https://mcp.fast.io/mcp/operations`
- Coding agents in IDEs (Cursor, Claude Code, Cline, VS Code agent mode) connect at `https://mcp.fast.io/mcp/code`

Through these MCP endpoints, an engineer using Cursor or a researcher using Claude can ask complex questions against the company brain while writing code or drafting strategic reports. The AI assistant queries the Fast.io workspace, retrieves verified passages with full citations, and incorporates that knowledge directly into ongoing work. Readers can review complete setup instructions in the [remote MCP server setup guide](https://mcp.fast.io/docs).

### Workspace Permissions and Activity Tracking

Enterprise knowledge bases require clear access boundaries. Fast.io enforces role-based permissions set per organization, workspace and share, by member role. There are no folder-level or file-level permissions; every member of a workspace accesses files according to their assigned role.

This model keeps permissions clean and predictable. Material intended only for executives, legal teams, or specific project groups is organized into its own dedicated workspace or share. Search and Intelligence mode query only from the specific workspaces and shares the asking person or agent belongs to, preventing unauthorized information retrieval. When files sync from Google Drive, Dropbox, Box, or OneDrive, they take the Fast.io workspace's permissions rather than mirroring source drive access lists.

Changes within the workspace are recorded in a detailed activity log. The log tracks uploads, edits, moves, deletions, membership updates, permission adjustments, and share creation across both people and AI agents.

### Structured Extraction with Metadata Views

Beyond conversational search, organizations often need to extract structured data from collections of documents. Fast.io includes [Metadata Views](/product/document-data-extraction/), which turn document collections into queryable, filterable spreadsheets.

Users describe the fields they want extracted using natural language, and the system creates typed schemas (Text, Integer, Decimal, Boolean, URL, JSON, Date & Time) across PDFs, scanned forms, and spreadsheets. Legal teams can extract contract renewal dates and counterparties, finance teams can capture invoice totals, and operations teams can track vendor compliance across hundreds of documents simultaneously without building manual extraction templates.

### Predictable Workspace Pricing

Rather than charging per-user subscription fees that escalate as headcount grows, Fast.io operates on flat workspace plans. Monthly plans start with a 14-day free trial, which requires a credit card.

The plan lineup scales predictably across team requirements:

| Plan Tier | Monthly Pricing | Annual Pricing | Included Seats | Included Storage & Limits | Included Monthly AI Credits |
| --- | --- | --- | --- | --- | --- |
| Starter | $9.99/mo | $99/yr | 3 seats | 250 GB storage (25 GB max upload) | 100,000 credits |
| Business | $49.99/mo | $499/yr | 10 seats | 5 TB storage (50 GB max upload) | 600,000 credits |
| Enterprise | $199.99/mo | $1,999/yr | 30 seats ($1/seat up to 200) | 25 TB storage (100 GB max upload) | 3,000,000 credits |
| Add-on Usage | Billed in currency | Flexible terms | Additional seats on Enterprise | Storage 1.5 cents/GB, bandwidth 4 cents/GB | Overage $10 per 100,000 credits |

Credits meter AI work such as semantic indexing and conversational queries, while storage and bandwidth beyond plan allocations are billed in currency rather than credits. Review complete tier specifications on the [Fast.io pricing](/pricing/) page. This flat tier structure allows companies to scale their knowledge base across dozens of team members and automated agents without facing steep per-seat licensing penalties.

## Frequently asked questions

### How much does Guru cost per month?

Guru does not publish prices today. Its pricing page presents quote-based packages built from the Platform, the Expertise, and the Infrastructure, scoped to organizational scale, knowledge complexity, and AI maturity, so buyers request a proposal from sales. Archived copies of the page show earlier published rates of $15 per user per month billed annually, or $18 billed monthly, in April 2024 and May 2025, and $25 per seat per month, or $30 billed monthly, with a 10-seat minimum in January 2026.

### Is there a free version of Guru knowledge base?

Guru does not provide a permanent free tier for its knowledge base platform. While the company previously offered a basic free starter tier for up to three users, that option was retired. Guru's current pricing page directs buyers to a call with its team, where Guru scopes a package.

### What is included in Guru All-in-one AI plan?

The All-in-one plan was the self-serve plan on Guru's pricing page in 2024 and 2025, listed at $15 per user per month billed annually or $18 billed monthly. It bundled AI search across connected sources, an intranet, and a wiki. Guru no longer lists it; its current pricing page offers quote-based packages instead.

### How does Guru meter AI answers and query credits?

Guru meters AI work in credits drawn from a shared pool for each workspace, and credits are not assigned to individual users. One search or one generated answer consumes one credit, and asking a Knowledge Agent to create a card, where it searches, drafts, and publishes, consumes three. Admins are notified as the workspace nears its allocation and can add credits.

### What does card maintenance cost in Guru?

Guru searches connected sources such as Google Drive, SharePoint, Slack, and Zendesk alongside its cards, and Knowledge Agents can draft and publish cards. The ongoing cost is verification. Every card has an assigned verifier, cards can come up for review every 30, 60, or 90 days or on a longer interval, and cards that restate source documents need an update and a fresh verification when those documents change.

### How do intelligent workspaces differ from card-based knowledge bases?

Intelligent workspaces index intact files like PDFs, Word documents, and spreadsheets directly upon arrival, using retrieval-augmented generation to provide citation-backed answers with specific page references. Card-based knowledge bases keep their trusted answers in cards that carry their own verifier and review interval, even when they also search connected sources.

## Sources

- [Guru: Pricing for our AI-Powered Knowledge Management Platform](https://www.getguru.com/pricing): Guru prices its current packages to each organization rather than publishing per-seat rates.
- [Guru Help Center: About Guru Credits](https://help.getguru.com/docs/about-guru-credits): Guru AI credits are a shared pool for each workspace rather than an allowance assigned to each user.

## About Fast.io

Fast.io provides shared workspaces where people and AI agents work on the same files, with built-in semantic search and citation-backed chat over what they hold. Agents reach it through a remote MCP server, a REST API at https://api.fast.io/current/, and a command line client published on npm as @vividengine/fastio-cli. MCP setup is at https://mcp.fast.io/docs: Claude and most MCP clients connect to https://mcp.fast.io/mcp/tools, ChatGPT to https://mcp.fast.io/mcp/operations, and coding agents to https://mcp.fast.io/mcp/code.