AI assistants
One assistant per need: role, tone, knowledge, tools and publishing rules, configured from a single interface and versioned.
Features
Everything you need for assistants that are useful and under control
Eleven functional areas, shown with their actual availability: what works right after installation, what requires a Pro license or configuration, and what is still to come.
Knowledge bases
Import your files, index websites, transcribe audio and video, and screen every document before it feeds into answers.
Sourced answers (RAG)
Hybrid search, contextual enrichment and verified citations: every answer points to the excerpts that support it.
Language models
OpenAI, Mistral, Anthropic, Gemini or a model you host yourself: a choice per assistant, encrypted keys and a benchmark to decide based on facts.
Publishing and channels
Widget on your website, public page, internal portal, email, Slack and phone: one assistant, many touchpoints.
Phone callbot
One number, one voice, the same knowledge as the chat: the callbot answers calls, routes them by topic and transfers to an agent when needed.
Tools and agents
Connect your services via MCP, webhooks or REST APIs: the assistant queries your systems, triggers actions and hands the conversation over to the right person.
Document workflows
Synchronize folders, FTP/SFTP servers and websites, have changes approved and publish knowledge base versions on a schedule.
Monitoring and improvement
Metrics, conversation review, AI analysis, reports, live takeover by an agent and a multichannel ticket center.
Security and governance
Roles, SSO including FranceConnect+, white-label organizations, GDPR retention, audit log, content validation and migration between environments.
API and developers
A documented REST API (OpenAPI), API keys, JSON responses with sources and real-time streaming.
Availability
The details, capability by capability
We clearly distinguish what is available from what depends on a condition. Each status is established from the platform’s source code.
- Available Feature implemented and usable with the standard configuration.
- Pro license Feature available with a Pro license.
- Requires setup Feature available subject to configuration, a key or a third-party service.
- Experimental Feature with limited scope or not yet validated in real-world conditions.
- Under consideration Planned improvement, not available today.
- Not available Feature not included in the platform.
AI assistants11 capabilities
| Capability | Status |
|---|---|
| AI-guided creation assistantIdentity, system prompt, welcome message and suggestions generated, then editable. | Available |
| System prompt generation and improvement | Available |
| Role and tone presets | Available |
| Response languageThe user’s language or an enforced language. | Available |
| Guided flow (decision tree)Multiple-choice questions before or during the conversation, JSON import/export. | Available |
| “Configurable” assistantForm fields and variables injected into the prompt for specialized assistants. | Available |
| Configuration versioningHistory, comparison, non-destructive restore, version comments. | Available |
| Duplication, activation, side-by-side comparison | Available |
| Rich content in responsesAction buttons, quick replies, carousel, images. | Available |
| Consent-based geolocationUser location via the Base Adresse Nationale (French national address database), with explicit consent. | Requires setup |
| Voice conversationPreview version, limited to signed-in users. | Experimental |
Knowledge bases11 capabilities
| Capability | Status |
|---|---|
| PDF, Word (.docx), PowerPoint (.pptx), Excel (.xlsx) documents | Available |
| Plain text, Markdown, CSV, HTML, JSON | Available |
| Batch upload20 files of 20 MB per upload by default, adjustable by the administrator. | Available |
| Website indexingStart URL, include/exclude patterns, depth, page count, linked PDFs, sitemap, scheduling. | Pro license |
| Audio and video transcriptionMP3, WAV, M4A, MP4, MOV, MKV… via Whisper (OpenAI) or Mistral. | Requires setup |
| OCR for images and scanned PDFsVia the Mistral OCR service (key required). | Requires setup |
| Content validation before indexingFile safety, sensitive content, personal data, prompt injection; quarantine and human decision. | Requires setup |
| Knowledge base versions with approvalProduced by workflows: comparison, approval, publication. | Requires setup |
| One-off attachment in a conversationAnalyzed for the conversation without being added to the knowledge base. | Available |
| SharePoint, Google Drive, OneDrive, Confluence, S3 connectorsNot available today; can be achieved by developing a connector. | Under consideration |
| .doc, .xls, .ppt and OpenDocument formats | Not available |
Sourced answers (RAG)9 capabilities
| Capability | Status |
|---|---|
| Hybrid semantic + full-text searchRank fusion (RRF), on PostgreSQL and pgvector. | Available |
| Contextual enrichment of excerptsEach excerpt is indexed with a sentence of context about its document. | Available |
| Multi-query rephrasing | Available |
| Knowledge base prioritiesFrom 1 to 10: in case of contradiction, the higher-priority source prevails. | Available |
| Verified citationsDocument name, excerpt, original URL for web pages, viewing permissions per knowledge base. | Available |
| “Insufficient evidence” warningShown when no relevant source was found. | Available |
| Deep searchDedicated mode, with a specific model. | Available |
| Model-based result rerankingAvailable when the internal model is OpenAI. | Requires setup |
| Complementary web searchVia the Serper service (key required). | Requires setup |
Language models8 capabilities
| Capability | Status |
|---|---|
| OpenAI, Mistral, Anthropic, Google Gemini | Available |
| Ollama, vLLM and OpenAI-compatible servicesSo answers can be generated by a model you host yourself. | Available |
| Live model list and connection test | Available |
| Extended reasoningThinking budget (Anthropic), reasoning effort (OpenAI). | Available |
| Structured JSON outputWith an optional JSON schema. | Available |
| Encrypted credential vault (AES-256-GCM)Sharing with named users, or availability with a daily cap. | Available |
| Multi-model and multi-version benchmarkTokens and latency per question, Excel export. | Available |
| Fully offline operationEmbedding, transcription and OCR currently go through OpenAI or Mistral. | Experimental |
Publishing and channels11 capabilities
| Capability | Status |
|---|---|
| Embeddable widget (script, iframe, inline)Label, color, size, auto-open, organization branding. | Available |
| Dedicated public pageCustomizable URL, choice of display theme. | Available |
| Employee chat portalHistory, Excel and PDF export, attachments, sharing. | Available |
| Answer rating and satisfactionThumbs up/down with comments, 1-to-10 rating at the end of the conversation. | Available |
| Public sharing of a conversationRead-only snapshot, expiration, revocation. | Available |
| Email channel (IMAP/SMTP)Automatic reply within the email thread. | Requires setup |
| Slack channelDirect messages and mentions in channels. | Requires setup |
| WhatsApp BusinessPreview version; Meta Business account required. | Experimental |
| Phone callbotThe assistant answers your phone lines: see the Phone callbot page. | Available |
| In-browser voice, AlexaEarly-stage developments, not validated in production. | Experimental |
| Microsoft Teams | Not available |
Phone callbot15 capabilities
| Capability | Status |
|---|---|
| Voice answers on your phone numbersSpeech recognition, assistant response, then speech synthesis, in real time. | Available |
| Same knowledge as the chatKnowledge bases, tools and rules of the linked assistant; short answers suited to speech. | Available |
| Voice barge-inThe caller can interrupt the assistant, which stops and listens. | Available |
| French and EnglishAutomatic detection of the caller’s language. | Available |
| Voice settings per lineSpeech recognition and synthesis provider, voice, speaking rate, pauses, domain vocabulary. | Available |
| Failover to a backup voice providerIf an outage occurs during a call, without dropping the call. | Requires setup |
| Routing by topic and business hoursTime zone, past-midnight schedules, public holidays. | Available |
| Transfer to an agentOn request, by pressing 0 or after repeated silences; a context summary is passed on, and the assistant drops off as soon as the agent picks up. | Available |
| Callback requestTicket sent to the relevant team with a call summary; SMS confirmation if an SMS gateway is connected. | Available |
| End-of-call satisfaction surveyRating from 1 to 5, by voice or keypad. | Available |
| Test call from the browserSimulated phone audio quality, latency measured at each exchange. | Available |
| Call trackingDuration, outcome, language, topic, latency and AI-generated summary. | Available |
| Live call takeoverCalls waiting for an agent appear in the conversation center. | Pro license |
| Crisis modeCrisis message played first, then redirection to a dedicated number. | Available |
| Fallback to your existing phone menu systemIf the platform does not respond, the call is routed to the line’s interactive voice response (IVR) system. | Available |
Tools and agents8 capabilities
| Capability | Status |
|---|---|
| Remote MCP server clientModel Context Protocol, Streamable HTTP transport, tool selection. | Pro license |
| Webhook toolsURL, method, headers, timeout, parameter schema. | Pro license |
| Declarative REST API toolsPath, query, header or body parameters, fixed by the administrator or left to the model; limited response size. | Pro license |
| Escalation to a humanContact form offered to the user, or live takeover by an agent. | Available |
| Structured data exportDownloadable tables generated by the assistant. | Available |
| Assistant treesVisual canvas and handoff of a conversation to a specialized assistant (up to 3 handoffs). | Pro license |
| Downloadable n8n workflow templateTo connect EasyRag to your n8n automations. | Requires setup |
| EasyRag exposed as an MCP serverNot available: EasyRag consumes MCP tools but does not publish any. | Not available |
Document workflows8 capabilities
| Capability | Status |
|---|---|
| Workflow creation assistant | Available |
| Visual step editorSource, sync, filter, validation, publication, condition, notification, end. | Available |
| Sources: local folder, FTP/FTPS, SFTP, websiteEncrypted secrets; SFTP host fingerprint recommended. | Requires setup |
| Knowledge base versionsDraft, pending, approved, published; comparison of additions, changes and deletions. | Available |
| Cron scheduling and time zones | Available |
| Webhook notification | Available |
| Transformation step (OCR, conversion, enrichment)Slot planned, not yet active. | Under consideration |
| Email notification | Under consideration |
Monitoring and improvement9 capabilities
| Capability | Status |
|---|---|
| Metrics dashboardFilters by assistant, model, language, topic and period. | Pro license |
| Conversation review and rating | Pro license |
| Request logsQuestion, answer, latency, tokens, sources used. | Pro license |
| AI analysis of conversationsTopics, quality, friction points, recommendations. | Pro license |
| Automatic topic classification and translation for supervisors | Pro license |
| Custom reports and PDF export | Pro license |
| Live takeover by a human agentExclusive handling, web push and email notifications. | Pro license |
| Conversation centerMultichannel tickets, teams, SLAs, assignment rules. | Pro license |
| Model benchmark | Available |
Security and governance11 capabilities
| Capability | Status |
|---|---|
| Roles: administrator, owner, editor, tester, support, viewer | Available |
| Accounts, email verification, password reset | Available |
| Google SSO and OpenID Connect (Keycloak, Azure AD…) | Requires setup |
| FranceConnect+ (eIDAS substantial or high)Authorization to be obtained from DINUM, the French government’s digital agency. | Requires setup |
| Organizations and white labelingLogo, colors, font, with WCAG contrast checking. | Available |
| Conversation retentionArchiving, anonymization or deletion after inactivity; dry run; legal hold. | Available |
| API call audit log and technical alerts | Available |
| Content validation before indexingPolicy per knowledge base, quarantine, traceable human decision. | Requires setup |
| Migration between instancesExport/import with dry run, report and rollback within 30 days; secrets are never exported. | Available |
| Crisis modePriority message added to responses, per organization or globally. | Available |
| OpenTelemetry, Prometheus and Grafana observability | Requires setup |
API and developers8 capabilities
| Capability | Status |
|---|---|
| Documented /api/v1 REST API (OpenAPI / Swagger) | Available |
| Complete answer in a single call (chat/complete)Text or JSON answer, sources, session and conversation IDs, usage. | Available |
| Streaming responses (Server-Sent Events) | Available |
| API keys with expiration | Available |
| Full management through the APIAssistants, knowledge bases, documents, tools… according to the key owner’s permissions. | Available |
| n8n (template provided) and Power AutomateVia the API, configured in the automation tool. | Requires setup |
| Make, ZapierNo native integration; possible through a generic HTTP call. | Not available |
| Keys with restricted scopes | Under consideration |
Want to see these features in action?
Try the demo assistant, or tell us about your project: we’ll get back to you with a concrete feasibility assessment.