For LegalTechs, law firms and compliance teams that need traceable answers on Uruguayan legislation and case law, without building or maintaining the corpus from scratch.
API, internal product or private deployment · We never train models on your queries
Use one, two or all three. Each layer is independent and integrates over a REST API with Bearer authentication.
Normalized Uruguayan legislation and case law, legally chunked and vectorized. Ready to ingest into your vector store.
Semantic + keyword search endpoint with hybrid ranking calibrated for the Uruguayan legal domain.
Answers generated with verifiable citations. Send the query, get back the answer with the exact sources that support it.
LegalBench-RAG research makes it clear: in law, what matters is retrieving minimal, highly relevant passages, not generic chunks. Bad chunking produces incorrect citations — and in the legal domain, that has consequences.
Every article, subsection and paragraph is an independent semantic unit. Naïve token chunking splits clauses and loses cross-references between articles.
Laws are amended, repealed and updated. Without versioning, your RAG may cite legislation that is no longer in force — the most critical problem in production.
Pure BM25 fails with legal terminology. A semantic-only index fails with case names and ruling numbers. Hybrid ranking calibrated for Uruguay solves both.
Every source is ingested from its official channel. No PDFs of unknown origin, no third-party copies.
Rulings from every chamber. Civil, Labor, Criminal and Family chambers.
TCA rulings on administrative acts and public law.
TAT, TAC, TAP, TAF — second-instance rulings by subject matter.
National Case Law Database. Access via the Judiciary’s official web services.
Complete national legislation since 1830. Open JSON data under IMPO license.
Bills, enacted laws and verbatim transcripts of sessions.
Executive decrees, ministerial resolutions and administrative acts.
BCU financial regulation · DGI tax criteria · ratified ILO conventions.
About public legislation: IMPO declares its database as open data available for reuse. What we sell is not "the raw law" — it is the normalized, enriched, versioned and legally chunked corpus, plus the specialized search and retrieval infrastructure. That processing is where the value is.
The point is not to generate convincing legal text. It is to retrieve the correct rule or ruling, show where it comes from, and leave an auditable trail so a lawyer can review it.
Generated answer: In principle, a unilateral salary reduction may constitute a serious breach by the employer and entitle the worker to claim constructive dismissal. The final outcome depends on the evidence, the size of the reduction, its permanence and the context of the employment relationship.
Ruling on unilateral modification of essential conditions of the employment contract.
Statutory reference related to salary, labor claims and recovery.
The same infrastructure is packaged differently per buyer: API for products, private demo for law firms and an internal solution for companies.
Add Uruguayan legal search and answers with citations inside your product without building your own scraping, embeddings, legislative updates or legal ranking.
Internal search over the public corpus and your own documents: precedents, contracts, briefs and frequent queries. Private deployment option for teams with sensitive data.
Internal queries for HR, compliance, finance or legal departments with answers grounded in current Uruguayan legislation and source traceability.
Complete corpus for legal NLP research, case-law analysis and academic studies. Export in Parquet or JSON for offline analysis.
AI-assisted due diligence and regulatory analysis. Search across DGI, BCU and sector regulations with auditable citations for client reports.
Embed a case-law search widget in your portal. Your users search rulings directly from your platform, with your branding.
Processed Uruguayan corpus, verifiable citations and integration designed for teams that need reliable answers, not approximations. Built for:
The corpus is normalized and legally chunked by article — not raw PDFs dumped into a vector store. Processed to answer in production, not in a demo.
Every result links to the official article or ruling. No unsupported answers: real traceability for reports, due diligence and compliance.
Search API and RAG-as-a-service over the Uruguayan corpus. Integrate in days with a REST API and SDKs, without building or maintaining the corpus yourself.
Standard REST API. SDKs for Python and JavaScript. Compatible with any stack.
Request access and we send your credentials within 24h. Standard Bearer token on every endpoint.
One POST to the search endpoint and you already get legal results with score and normalized metadata.
Rate limits configurable per plan. Webhooks for corpus updates. Monitoring dashboard included.
Built from day one to meet the standards required by law firms, fintechs and companies with sensitive data.
Each client operates in a fully isolated namespace. No data, query or result crosses between accounts. Multi-tenant architecture with database-level separation.
TLS 1.3 on every endpoint. Vectorized data and indexed corpus encrypted at rest with AES-256. API keys are never stored in plain text.
The queries you send to the API are not stored, not logged with content and never used to train or fine-tune models. Configurable retention period (default: 0 days).
Guaranteed uptime with a public status page. The Enterprise plan includes an SLA with financial penalties for breaches and on-call technical support.
The corpus and the API can run in your private infrastructure or VPC. No dependence on our servers, no data traffic outside your environment. Available on the Enterprise plan.
Every corpus fragment has full traceability: official source, ingestion date, version and validity status. You can audit any citation back to its origin.
Data residency: By default, data is processed on infrastructure located in the region. For Enterprise customers with specific data sovereignty requirements, we offer deployment in Uruguay or the SA region. Contact us to assess your case.
Pay for what you use. No volume commitments on the first two tiers.
To evaluate the API and build a POC. No credit card.
For production products that need the full corpus and RAG.
Private deployment, isolated namespace, guaranteed SLA and dedicated support.
The answers that business, legal and tech teams usually need before moving forward with a pilot.
No. It is designed as search, retrieval and assistance infrastructure. The professional decision still belongs to the lawyer or legal team. That is why every answer includes sources, citations and traceability.
The system is designed to reduce that risk with grounding, hybrid ranking and mandatory citations. If it does not retrieve enough sources, it can be configured not to answer. For critical flows we recommend human review.
Yes. The Starter plan allows a technical POC at no cost. For law firms or companies we can arrange a private demo with real team cases and a limited scope.
Yes. For non-technical teams we package it as a private demo, internal search tool or assisted integration. The API is one way to integrate, not the only way to use the corpus.
Yes, on Pro and Enterprise plans. You can upload contracts, internal precedents, manuals or briefs to your namespace. Private documents are never accessible from other accounts.
Yes, per API key. Starter: 10 req/min, 500/mo. Pro: 100 req/min, 20,000/mo. Enterprise: configurable. Response headers include X-RateLimit-Remaining and X-RateLimit-Reset. 429 responses include Retry-After.
Partially. The /v1/complete endpoint follows a structure similar to chat/completions but with additional fields: citations[], sources[] and grounding_score. There is an openai_compat: true mode that omits those extra fields if you already have an OpenAI client configured.
Yes, on Pro and Enterprise plans. You configure filters (by source, subject, keywords) and receive a POST to your endpoint when new matching content arrives. Useful for keeping a local index synced or triggering automatic alerts.
On Starter, the overage blocks until the next month. On Pro and Enterprise, overage is billed at a per-query rate agreed in advance — we never cut off service without notice. You can set up automatic alerts when you reach 80% of your quota.
Official SDK for Python (pip install abogacia-py) and JavaScript/TypeScript (npm install @abogacia/client). Both include types, error handling, automatic retries and streaming support. For other languages, the API is standard REST — cURL works perfectly.
Request API access, a private demo or a limited pilot. We reply with the best path for your case: product, law firm, compliance or research.
We reply in under 24 business hours.
We will contact you in under 24 business hours with your access credentials and next steps.