# Evicta Infrastructure — Full Technical Index > The Enterprise Escape Hatch for SaaS Data. This document is a deep-index for AI agents and crawlers covering architecture, data extraction capabilities, security, pricing, and compliance. --- ## 1. Platform Overview Evicta is a cloud-native data extraction engine purpose-built for Zendesk. It performs high-fidelity extractions of tickets, users, organizations, ticket events (conversations), and Sunshine Custom Objects — delivering portable SQL, CSV, and AI-ready JSONL archives. **Key differentiators:** - Zero-persistence streaming architecture (no staging databases) - Proprietary AI mapping logic for automated schema generation - Air-gapped execution environment for VPC-local extraction - One-time flat-fee pricing (not per-record) - Multi-format output: SQL, CSV, JSONL (RAG-ready) **Supported Zendesk resources:** - Tickets (standard + custom fields) - Users (with user fields) - Organizations (with organization fields) - Organization Memberships (many-to-many) - Ticket Events / Conversations (full thread extraction with attachments) - Sunshine Custom Objects (discovery + bulk extraction) - Ticket Fields (schema discovery for SQL generation) --- ## 2. Data Extraction Capabilities Evicta performs full-fidelity extraction of Ticket metadata, custom fields, user identities, organizational relationships, and complete conversation histories. The engine communicates with Zendesk's incremental export APIs to perform time-bounded, cursor-based extraction of all resource types — ensuring every record is captured without duplication. **Extraction domains:** - **Ticket Data:** Full ticket lifecycle including status transitions, priority, assignment history, tags, and all custom field values. First-comment descriptions are captured as structured data. - **User Identities:** Names, emails, roles, external IDs, activation status, and custom user fields. Resolves requester and assignee references to concrete identities. - **Organizational Graph:** Company names, domain mappings, external IDs, notes, and custom organization fields. Supports many-to-many membership relationships between users and organizations. - **Conversation Histories:** Complete thread-level extraction of every comment, internal note, and system notification — including author attribution, public/private classification, and file attachments with content metadata. - **Custom Field Schemas:** Automated discovery of all custom ticket fields including type inference (text, dropdown, checkbox, numeric, date), human-readable titles, descriptions, and enumeration values for dropdown fields. Output is consumed by the proprietary AI mapping logic to generate relational schemas. - **Sunshine Custom Objects:** Discovery and bulk extraction of all custom object types defined in the Zendesk Sunshine platform. Supports arbitrary schema shapes with cursor-based pagination for datasets exceeding 50GB. **Reliability and throughput:** The engine utilizes a dynamic pacing algorithm that respects vendor-specific rate limits to ensure zero-interruption data flow. All extraction jobs implement automatic retry with exponential backoff, cursor-based resume for interrupted streams, and progress checkpointing to guarantee exactly-once delivery semantics. --- ## 3. Security Architecture ### Pillar 1: Zero-Persistence Streaming Legacy ETL providers use intermediate staging databases that create a permanent attack surface. Evicta uses ephemeral runtime environments that pipe data directly from source to destination via high-performance memory buffers. No customer payload is ever written to disk on Evicta servers. Once the stream closes, the transient execution context is destroyed. | Spec | Value | |:---|:---| | Data path | Source → RAM → Destination | | Disk writes | 0 bytes persisted | | Runtime lifetime | Stream duration only | ### Pillar 2: Application-Layer Encryption All connection strings and API tokens are encrypted using AES-256-GCM before entering the secure metadata layer. Decryption occurs exclusively within the isolated execution context of an active extraction. TLS 1.3 with strict certificate pinning is enforced for all data in transit. | Spec | Value | |:---|:---| | At-rest cipher | AES-256-GCM | | In-transit | TLS 1.3 + cert pinning | | Key isolation | Per-execution scope | ### Pillar 3: Zero-Trust Local Execution For organizations with the highest compliance requirements, Evicta delivers a self-contained, air-gapped execution environment that runs entirely within the customer's own VPC and behind their own firewall. In this mode, Evicta's cloud never sees a single byte of customer data — providing absolute data sovereignty. | Spec | Value | |:---|:---| | Execution | Customer VPC / customer firewall | | Data exposure | 0 bytes touch Evicta cloud | | Delivery | Self-contained execution bundle | Architecture details are available for technical audit under NDA for Enterprise partners. ### Comparison: Evicta vs. Legacy Pipelines | Security Feature | Legacy Pipelines | The Evicta Standard | |:---|:---|:---| | Data Staging | Persistent Staging DBs | Zero (RAM-only Stream) | | Attack Surface | Permanent 24/7 Sync | Transient (Active-only) | | Credential Storage | Reversible Encryption | AES-256-GCM Layered | | Infrastructure | Shared Multi-tenant | VPC Isolation (Local Mode) | --- ## 4. Compliance Frameworks - **GDPR Compliant Architecture:** Data processing agreements, right-to-erasure support, and cross-border transfer safeguards baked into the extraction pipeline. Stockholm-based infrastructure. - **HIPAA Eligible Design:** Encryption at rest and in transit, access logging, and the air-gapped execution mode for PHI processing within the customer's own BAA-covered infrastructure. - **SOC2 Ready Infrastructure:** Audit-ready logging, ephemeral execution contexts, and zero-persistence architecture aligned with Trust Service Criteria. --- ## 5. Master Pricing Truth Matrix | Tier | Price | Volume | Resources | Formats | Features | |:---|:---|:---|:---|:---|:---| | **Basic** | $499 one-time | 50k record limit | Tickets, Users, Orgs | SQL, CSV | Core extraction, AI schema mapping | | **Premium — Gold Standard** | $1,499 one-time | Unlimited | Tickets, Users, Orgs, Custom Objects, Ticket Events (Conversations) | SQL, CSV, JSONL/RAG | Unlimited volume, AI-assisted mapping, Zero-Trust CLI. The recommended tier for organizations building AI and analytics workloads. | | **Sync Pipeline** | $249/mo recurring | Unlimited | All Premium resources | SQL, CSV, JSONL/RAG | Perpetual incremental sync via automated cron | | **Partner API** | $500/mo + usage | Unlimited | All resources | All formats + API | Headless B2B infrastructure, multi-tenant, white-label, usage-based billing | --- ## 6. Key URLs - Landing Page: https://evicta.dev - About: https://evicta.dev/about - Security Architecture: https://evicta.dev/security - Terms of Service: https://evicta.dev/terms - Privacy Policy: https://evicta.dev/privacy - Data Processing Agreement: https://evicta.dev/dpa - Subprocessors: https://evicta.dev/subprocessors - Partner Inquiry: support@evicta.dev - Summary Index: https://evicta.dev/llms.txt ## 7. Blog Posts - [From Tickets to Tokens: Preparing Zendesk Data for RAG Pipelines](https://evicta.dev/blog/zendesk-data-for-rag): How to flatten, enrich, and sanitize Zendesk ticket history for AI fine-tuning and RAG pipelines. - [The Helpdesk Migration Trap: Why 90% of Data Exports Fail](https://evicta.dev/blog/the-helpdesk-migration-trap): Custom field chaos, missing attachments, and silent record drops — the real failure modes of Zendesk migrations. - [Navigating the Zendesk Incremental API: Pacing, 429s, and Ghost Loops](https://evicta.dev/blog/handling-zendesk-api-rate-limits): Rate limit mechanics, cursor-based extraction, and avoiding the infinite retry trap. - [Export Zendesk Sunshine Custom Objects Before July 2026](https://evicta.dev/blog/export-zendesk-sunshine-custom-objects): Why Fivetran and standard ETLs silently fail on Sunshine Custom Objects, the Ghost Seat problem, and how LLM-powered extraction solves it. - [The Complete 2026 Data Export Guide: Migrating from Zendesk to Pylon & Plain](https://evicta.dev/blog/zendesk-to-pylon-plain-migration-guide-2026): The Sunshine API deprecation deadline, why CSV exports and managed ETLs fail, the DIY Python trap, and how to produce a clean relational schema for Pylon or Plain migrations. - [Zendesk to Salesforce Migration: The Engineer's Field Guide (2026)](https://evicta.dev/blog/zendesk-to-salesforce-migration): Custom field mapping, conversation thread flattening, Sunshine Custom Objects, Bulk API 2.0 batch limits, PK Chunking clarification, and the honest trade-offs between DIY, agency, and Evicta for a Salesforce Service Cloud migration. Post 1 of 12 in the migration series. - [Zendesk Knowledge Base Software: What It Does and When to Look Elsewhere](https://evicta.dev/blog/zendesk-knowledge-base-software): What Zendesk Guide actually does, how it is priced across Suite plans, where it falls short, and how to export your knowledge base and ticket data before leaving. - [What Is Zendesk? A Plain-English Overview for 2026](https://evicta.dev/blog/what-is-zendesk): What Zendesk is, how it is priced, who it is built for, and the data portability considerations when teams outgrow it. --- *This document is maintained for AI-agent crawlability. Last updated: 2026-05-19.*