Skip to content
sacs.ai

California school finance / Agent ready

Your agent.
California
context.

The school-finance knowledge your AI is missing. Bring official SACS data, district context, and traceable evidence into the agent you already use.

Free public connector. Read-only tools. No SACS account needed.

A little context. A better starting point.
Your workspaceThe agent you already useA compatible remote MCP client
Model Context Protocol
sacs.aiContext layer

California school-finance context,
ready for the next question.

District identitySACS definitionsFinancial contextSource evidence
Official public CDE / SACS datasets
Context in. Evidence stays attached.
Built for the work ofDistrict finance teamsSchool business consultantsCounty offices of education

01 / The missing context

A capable agent.
A more informed conversation.

California school finance has its own language. SACS.ai brings the district, the reporting basis, and the source into view.

01 /

Find the right district.

Resolve district identity and discover official datasets before asking a financial question.

District lookup / dataset discovery
02 /

Give the numbers context.

Bring SACS code definitions, financial summaries, and comparisons into the conversation.

SACS ontology / public financials
03 /

Keep the evidence attached.

Get compact district context packs with dataset references, provenance, and explicit limitations.

Single district / bounded batches

02 / Put it to work

Start with a district.
Ask a better question.

A useful first task: ask your agent for a briefing that keeps the numbers and their limitations together.

Try this in your connected agent

Use SACS.ai to prepare a district context pack for Los Angeles Unified. Resolve the district and discover available filings first. Select the latest available year and name the reporting basis. Request a summary for that year with fund_codes=["01"] for General Fund only. Include dataset references, source links where available, and limitations. Keep budgets, interims, and actuals distinct. Flag unavailable actuals and quarantined rows; do not fill gaps with estimates.

What to look for in the responseIllustrative

A district briefing
you can inspect.

An example structure, not a live data result.

01 / Identity
District name and identifiers, resolved before analysis.
02 / Reporting context
Fiscal year, reporting period, and basis kept explicit. A budget is not an actual.
03 / Evidence
Dataset references and versions, with user-openable source links where available.
04 / Limitations
Unavailable actuals stay marked unavailable. Quarantined rows stay out of conclusions.
Ask your agent to show its sources before you use the answer.

03 / Bring your own agent

One connection.
Your existing workflow.

MCP lets an AI agent call SACS.ai’s public tools when it needs California school-finance context.

Use a client that supports remote MCP over Streamable HTTP. Client setup and connector availability vary.

Open the agent quickstart: exact calls, use cases, and troubleshooting ↗ · Developer reference

Public MCP endpoint

https://mcp.sacs.ai/mcp

Transport: Streamable HTTP · Authentication: None

Access: Public financial data is read-only; optional registration and feedback require consent.

  1. 01

    Add a remote MCP connection

    In your agent’s connector settings, add SACS.ai using the endpoint above. This public connector does not require a SACS login or API key.

  2. 02

    Enable the SACS tools

    Allow your agent to use the connector in your conversation. Start with public district data.

  3. 03

    Try the district briefing

    Paste the first-use prompt, then check the reporting period, sources, and limitations in the answer.

Connection not working?

Check that your client supports remote Streamable HTTP and that the URL includes /mcp. Opening the endpoint in a browser is not a connection test. If a request is rate-limited, wait before trying again; for timeouts, try one district and a narrower period.

Still stuck? Email the client name and error to support. Individual client walkthroughs are still being validated.

04 / Bring your organization’s work

Public context.
Private working room.

For districts, county offices of education, CPA firms and consultants. Bring your team’s data and workflows to the agent you already use, with a web workspace for review.

Organization previewPremium features are being validated in staging. Paid production access has not launched.

Private datasets

Bring your DAT exports, inspect fund and resource statements, and share datasets within your organization.

Dataset checks

Run queued TRC checks and inspect saved reports. The migrated rule catalog is partial; SACS Web validation remains essential.

Versioned projections

Give your agent assumptions to preview and apply. Keep source data intact, check each revision, and export a DAT with its verification manifest.

Variance analysis

Compare budget, interim or actual datasets by account slices. Focus audit analytics with dollar and percentage thresholds, separately or together.

District watchlists

Track new official datasets in SACS.ai through your agent or the web. Opt-in daily email sends only when matching changes are available.

Organization controls

Manage members, roles, scoped agent keys, usage and your subscription. Use one shared workspace for people and agents.

SACS.ai Free

Start with public data.

$0

Official public financials, district discovery, SACS code definitions, comparisons and searchable CSAM guidance.

No SACS account required. Connect your compatible agent and ask a focused question.

Connect the free MCP

SACS.ai Organization

One subscription. Your whole team.

$4,800 / organization / year

$400/month equivalent, billed annually. Planned launch pricing.

  • Unlimited members and agents; shared service and compute limits apply.
  • Private datasets, checks, projections, variance analysis and watchlists.
  • The same organization price for districts, COEs and CPA firms.
Ask about Organization

Payments initially use Stripe only. Need a Purchase Order? Implementation, integrations and consulting are scoped separately. Your agent provider’s fees are separate.

Built to be inspected

Know what’s behind
the answer.

Sources, versions, and limitations belong in the work from the beginning.

How does an agent access my organization’s data?

Sign in to the web workspace and create a scoped API key for the organization. Use that key with the authenticated organization MCP connection shown in the workspace. The free public endpoint does not expose private datasets. You can revoke keys and manage member access from your organization.

Projection changes use preview and commit steps, with a separate report for each checked revision. Review your agent’s proposed changes and warning choices before using an exported DAT.

Is Organization priced per user or per district?

One annual subscription covers the subscribing organization’s members and agents, including a COE or CPA firm working across multiple districts. Shared service and compute limits apply. Your AI client or model provider may charge separately; those fees are not included in SACS.ai.

Can my agency pay by Purchase Order?

Premium subscriptions initially accept payment through Stripe only. Keeping billing simple helps us keep operating costs low and pass the savings on to users. We do not currently accept Purchase Orders or offer manual invoicing.

If your agency needs a Purchase Order, join the notification list from your organization’s Subscription page. Your interest helps us gauge demand before adding the cost of another payment method. Joining is free, does not start a subscription, and does not promise a launch date.

What data can my agent access?

The public connector exposes bounded, read-only tools over official public CDE/SACS data. It does not provide private uploads, unrestricted SQL, or access to your local files. Coverage varies by district, year, reporting period, and basis; ask for available datasets first.

How current is the context?

Reporting releases and imports happen on different schedules. SACS.ai does not promise exhaustive or real-time coverage. Check the returned dataset version, reporting basis, and limitations on each request. A recent retrieval timestamp does not mean a new reporting cycle is available.

How should I use the evidence?

Ask your agent to use sacs_get_dataset_source when offered. Published sources include the archived original file, its SHA-256, any separately labeled converted DAT, and a SACS Web viewer link with filing details. Older sources may not yet have published links. Distinguish reported figures from your agent’s interpretation and use professional review for consequential decisions.

What happens to my information?

The public tools are for public-data questions. Operational logs record bounded request and tool metadata, not prompts or returned financial rows by default. Your chosen agent provider has its own data policies. Do not send confidential documents through this public workflow.

Help shape what comes next

Put it to work.
Tell us what’s missing.

Useful, incomplete, incorrect, or stale? Share the district, reporting period, and request or data version if available. Public information only.

Share feedback by email