Perform queries and build AI agents in your AI-enabled workflow with confidence – knowing responses are grounded in ICE's proprietary fixed income evaluations: a consistent, traceable source.
Data available via ICE Data Services MCP connector
End-of-Day Fixed Income Evaluations
Daily evaluations for over 3 million financial instruments, representing over 150 countries and over 80 different currencies.
Enhanced Evaluation Transparency (EET)
160 data points per identifier across all EOD evaluations, basic reference data, price movement commentary, MBS assumptive data, and trade/quote info.
ICE AAA Municipal Bond Curve
Continuously updated transaction-driven curve used in ICE’s evaluation process, 15+ years history, rules based and fully transparent.
US Treasury Benchmark data
On-the-run U.S. Treasury yields used in evaluating U.S. Corporate and other similar bonds whose prices are spread-based.
FINRA TRACE & MSRB RTRS data
Bond trades across corporate, agency and securitized products reported to FINRA TRACE and muni bond trades reported to MSRB RTRS.
Use case examples and prompts
IPV, Compliance, Audit, Governance and many more functions can use this ICE data in their LLM and AI tools to support workflows for fixed income price verification, narrative generation, and portfolio-level analysis.
Task or query example 1:
Get day-over-day evaluation movement data that indicates whether the movement is idiosyncratic or systemic
Sample prompts
"Generate an audit narrative for today's evaluation of ISIN US017175AD24."
Task or query example 2:
Discover bond movement drivers across overall market, sector, and rate dynamics:
Sample prompts
"Were there any block prints that may have moved the evaluation?"
Task or query example 3:
Generate and view aggregated, portfolio-level day-over-day evaluation changes, decomposed into rate vs. spread, with sector bond type and maturity-bucket breakdowns:
Sample prompts
"Summarize last week's evaluation changes for my MBS holdings."
Task or query example 4:
Generate evaluation exceptions, overrides, and signoff data to support pricing controls and validation workflows:
Sample prompts
"Explain the prepayment assumption change in my GNMA MBS holding."
Price evaluations data with integrity intact
ICE's fixed income data is delivered via an MCP connector with the underlying asset-class evaluation methodology and context mapping built in. This means queries are grounded in a consistent, auditable data foundation, giving AI tools the structured context they need to reason over fixed income instruments effectively.
Methodology-bound knowledge layer
ICE's Fixed Income evaluations methodology is embedded as a knowledge layer provided to the MCP connector. This means methodology is applied as relevant with each query.
An asset class router maps each security to its proper evaluation frame, so LLM reasoning and responses can use the relevant ICE framework.*
Cross-layer data orchestration
As a bond's day-over-day price can move at multiple levels of the market simultaneously ICE provides the MCP connector structured access to relevant layers so the LLM can perform cross-layer reasoning and responses efficiently.*
- Rate level: US Treasury & AAA Muni curves.
- Issuer level: Issuer reference & curves.
- Bond level: ICE Evaluations & reason codes
*LLM reasoning and responses vary. ICE does not control nor is responsible for third party LLM outputs and use.
ICE Aurora
Bringing AI into what we do comes naturally given that technology and innovation have been foundational to ICE since our inception over 25 years ago.
Our approach is a natural extension of that legacy. Our goal with AI is to further enable automation for our customers, build tools that drive efficiency and deliver enhanced analytics and insights, all guided by rigorous governance standards and our commitment to the responsible use of AI.
Informing your AI workflow with ICE’s proprietary, structured fixed income data
Chris Edmonds, President, Fixed Income & Data Services, ICE
ICE connector for Claude
Connecting the ICE connector for Claude takes just a few steps.
- Open Claude Settings: Navigate to Settings / Connectors in Claude Desktop or Claude.ai
- Find ICE Data Services or Search for "ICE Data Services" in the Claude Connectors Directory
- Click Connect: You will be redirected to ICE SSO to authenticate
- Sign in with ICE SSO: Enter your ICE Data Services credentials and complete multi-factor authentication (MFA)
- Start querying - Once connected, try asking Claude. “Explain today’s evaluated price movement for this bond”
Requirements
- An active ICE Data Services account with applicable licenses to the available content sets in the MCP server
- Multi-factor authentication (MFA) enabled on your ICE SSO account
Search tips
Bring ICE’s quality fixed income data into your AI workflows with a fast, structured MCP connector.
Date ranges
- Use single-day or weekly date ranges for the fastest query performance. Queries spanning more than one month may time out.
- If no date range is specified, the system defaults to the last 7 calendar days.
- Data coverage begins January 2026. Earlier dates will return sparse or no results.
Identifying bonds
- Reference bonds by issuer name for convenient lookups. The MCP connector resolves issuers to the correct identifiers behind the scenes.
- You can also search by asset category, sector, maturity range, or other bond characteristics to find the securities you need.
Narrowing results
- Specify an asset category (corporate, municipal, agency, securitized) to focus your query and improve performance.
- Use the sector filter (Financial, Industrial, Utility) to narrow corporate bond queries.
- Use aggregation (group by sector, issuer, date, or hour) for volume analysis instead of requesting thousands of individual records.
Getting the most from Claude
- Be specific about what you want: "top 10 trades by size today" is better than "show me some trades."
- Ask follow-up questions — Claude maintains context about your previous queries.
- Use the knowledge layer for definitions: ask "What is an institutional trade?" rather than assuming thresholds.
ICE Data Services MCP connector – Data privacy
- Access to ICE data is governed by customers' existing ICE entitlements.
- ICE does not store query or session data beyond the duration of your connection with the connector.
- For data handling by your AI platform, refer to your AI provider’s privacy and data retention policies.
- For more information about ICE privacy and security
Have a question?
For questions, issues or feedback about our MCP connector please get in touch.