# Serasa Recomenda MCP for credit decisions. AI Agent Connect > Serasa Recomenda is a decision-as-a-service MCP that handles credit granting for CPF and CNPJ. It provides recommended limits, risk levels, and sales orientations across nine different distribution channels. Your AI client can automatically handle schema mismatches by dropping unsupported fields, ensuring requests always go through. ## Overview - **Category:** finance - **Price:** Free - **Endpoint:** https://edge.vinkius.com/vk_preview_XcVWWyaj401kLmDr2DDTDhVPzf7XeszOtGPZNaCf/ai-agent-connect - **Tags:** serasa, experian, credit-limit, credit-decision, brazil, pf, pj, recommendation ## Description You can use this MCP to turn Serasa Experian's credit data into actionable decisions. Instead of manually checking risk, you tell your AI client to submit a CPF or CNPJ along with the operation amount. The MCP handles the heavy lifting across nine specific distribution channels, including ecommerce, erp, and direct. It automatically cleans up your request by dropping any optional fields that don't fit the specific channel you've chosen, so you don't have to worry about schema errors. You can also request extra data like estimated revenue or credit scores through additional information modules. Once a recommendation is made, you can pull up specific proposals, list your history, or export everything into a clean report. It's built to work with OAuth2 authentication, making the connection to your agent straightforward. ## Tools ### generate_recommendations_report This tool exports a report of PJ recommendations. You can filter the results by name, document, proposal number, user, or status. ### get_proposal This tool fetches a specific recommendation using its proposal number. You just need to provide the same channel and document type used during the initial request. ### list_proposals This tool lists previously generated credit recommendations. You can filter by name, document, status, or a specific date range. ### recommend_credit_limit This tool generates a credit limit recommendation and risk level for a CPF or CNPJ. It supports various distribution channels and optional enrichment modules like score or estimated income. ## Prompt Examples **Prompt:** ``` Recommend a credit limit for CNPJ 33014556000196 for a 1500 operation on the direct channel, recurring customer. ``` **Response:** ``` The MCP calls recommend_credit_limit with the direct channel and returns the proposal number, recommended limit, risk level, and sales orientation. ``` **Prompt:** ``` What limit would Serasa suggest for this new customer, including score and negative annotations? ``` **Response:** ``` The MCP calls recommend_credit_limit with the additional information module for annotations, returning the recommendation along with the requested score and annotation data. ``` **Prompt:** ``` List the proposals we generated for that CNPJ through the distributor-prorede channel. ``` **Response:** ``` The MCP calls list_proposals using the distributor_prorede channel and the specific CNPJ to return the matching history. ``` ## Capabilities ### Automated Risk Assessment Your agent calculates recommended credit limits and risk levels based on CPF or CNPJ data. ### Multi-Channel Support The MCP routes requests through nine different distribution channels like ecommerce or erp. ### Data Enrichment You can request extra modules like estimated revenue or credit scores during the recommendation process. ### Proposal Management Your AI client can list, retrieve, or export credit recommendation history. ### Schema Error Handling The MCP automatically drops unsupported fields to prevent request failures. ## Use Cases ### E-commerce Credit Checks An online store uses the ecommerce channel to get instant credit limits for customers at checkout. ### ERP Integration A business automates credit decisions directly within their ERP workflow using the erp channel. ### Bulk Reporting A risk team generates a full report of all PJ recommendations to audit recent credit decisions. ### Customer Re-evaluation A lender checks the history of a recurring customer to adjust credit limits based on previous proposals. ## Benefits - Reduces request failures by automatically stripping incompatible fields for specific channels. - Provides instant risk levels and sales orientations for every credit request. - Centralizes credit decision history through searchable proposal lists and reports. - Enables deep data dives with optional modules for income and credit scores. ## How It Works Connecting to Serasa's decision engine is a direct process through Vinkius. 1. Connect your AI client to the Serasa Recomenda MCP via Vinkius. 2. Provide your OAuth2 credentials from the Serasa developer portal. 3. Instruct your agent to run a recommendation using a CPF, CNPJ, and amount. 4. Receive the limit, risk level, and proposal number immediately. ## Frequently Asked Questions **Which channels does this MCP support?** It supports nine channels: direct, distribuidor, distribuidor_prorede, ecommerce, ecommerce_relatorio, erp, erp_relatorio, consumo_externo, and large. **How do I handle errors with unsupported fields?** The MCP automatically drops any optional fields that are not supported by your chosen channel, preventing schema mismatch errors. **Can I get extra data like estimated income?** Yes, you can request additional modules like estimated revenue or income by using the informacoes_adicionais parameter. **How do I find a specific recommendation later?** You can use the list_proposals tool to find the proposal number or use get_proposal if you already have the ID. **What authentication is required?** The MCP uses OAuth2 client credentials, which you can obtain from the Serasa Experian developer portal. **Which channel should I use?** The channel mirrors how the credit is sold: direct for your own sales team, distribuidor or distribuidor_prorede for a dealer network, ecommerce / ecommerce_relatorio for online checkout, erp / erp_relatorio when the request comes from an ERP, consumo_externo for third-party consumption and large for high-value operations. **Why was my optional field ignored?** Each channel's schema accepts only a subset of the optional fields. The connector drops whatever the chosen channel does not support — for example parceiro is only valid on consumo_externo and large, and documento_indireto only on distribuidor_prorede — so the request stays valid. **What does cod_tipo_venda mean?** It marks the sale type: 1 for a new customer and 2 for a recurring customer. Combined with politica (1 or 2, the risk policy version) it selects which decision policy Serasa applies to the proposal.