citra-decision-system
Citra Decision System — agent learns from human judgement and recommends, human-approved decisions grounded in learning, SOPs and data and governed...
Tool overview
Overview
Citra Decision System is one of two open-source systems from Citra. It is self-hosted, runs with open models, and is designed so that data never leaves the network. It can be deployed with Docker Compose on your servers and can bring your own vLLM. It is Apache-2.0 on GitHub. The system handles decision work that needs a person: every case arrives assembled and scored, with a recommendation cited to the user's SOP, records and past decisions. An officer approves or overrides, and the system asks why. Corrections repeated by three different officers harden into a named, reversible judgement that shapes the next case. The agent never writes to systems on its own. Governance
features
include ontology-bound permissions, no AI-written SQL, display equals commit, and an auditable ledger. Evidence, recommendation, citations, decision, override and reason are stored in a tamper-evident, append-only ledger inside the user's own database. Fraud screening is off by default. The Community edition is Apache-2.0 and free forever, including the complete decision engine, builder, ontology, ledger and memory, maker-checker, approval queues, row filters, full memory curation, one team, one deployment, on your own infrastructure, and Citra Flows entire. Citra Enterprise is by quotation and adds an organisation-wide registry, retention and erasure policies, drift detection, champion-challenger runs, many teams/apps/deployments as one estate, and dedicated private cloud or your own with contracted SLA and roadmap priority.
Features
- AI assistant
- Natural language interface
- Workflow automation
Tags
Use Cases
Retail lending teams use Citra Decision System to assemble and score loan cases and cite the policy clause behind a recommendation.
Credit officers use Citra Decision System to approve, decline, or send a case for verification.
Operations teams use Citra Decision System to capture officer overrides and reasons for each decision.
Risk and audit teams use Citra Decision System to keep an append-only ledger of evidence, recommendations, citations, decisions, overrides and reasons.
Compliance teams use Citra Decision System to enforce ontology-bound read, build and write-back permissions.
IT teams use Citra Decision System to deploy on their own infrastructure with Docker Compose and bring their own vLLM.
Teams use Citra Decision System to work with Indian financial identifiers such as ₹, PAN, IFSC, GSTIN and Aadhaar.
Teams use Citra Decision System to turn repeated corrections by distinct officers into learned judgements applied to matching cases.
User Reviews
Related Tools
View moreLynote
Lynote combines AI-draft rewriting and AI-text detection with note-taking, transcription, video summaries, and study tools.
LMSYS Chatbot Arena Leaderboard
LMSYS Chatbot Arena is a crowdsourced open platform for LLM evals. Collected over 1,000,000 human pairwise comparisons to rank LLMs with the Bradley-Terry model and display the model ratings in Elo-scale.
智谱清言
Zhipu Qingyan is a GLM-based AI assistant whose official description emphasizes understanding goals, breaking down tasks, and using tools.
豆包
字节跳动AI助手