Run and adapt models
Choose the models, their runtime environment and the data entrusted to them.
Open source, deployable models and business integrations. We favour auditable architectures, chosen hosting and documented options for taking over a system.
We favour open-source software and solutions that can be deployed in an environment chosen with you. These building blocks can be used according to the project, with their roles explained and direct links to their sources.
Choose the models, their runtime environment and the data entrusted to them.
Keep data usable, formats documented and migration paths possible.
Define accounts, permissions and sharing in an environment you can administer.
Document deployment so it can be understood, rebuilt and developed further.
Make service health and useful operational signals visible.
Open languages and frameworks for products a team can understand.
The listed licences apply to the named projects or editions. Downloaded models, weights, extensions and cloud services have their own terms. Self-hosted software also requires updates, backups and ongoing operations.
A server’s location alone does not determine control over a system. We also examine operators, administrative access, subprocessors, outgoing data flows and recovery arrangements.
A dedicated server, an internal environment, hosting in France or in the European Union: the choice follows the project’s constraints.
Accounts, permissions, keys, logs and administration responsibilities are defined from the design stage.
Code, export formats, documentation and recovery procedures help prepare for changes or a different provider.
Architecture and data flows, licences, access management, hosting arrangements, backup and recovery, monitoring, documentation and portability.
Our approach to architectureThese groups illustrate possible connections according to your tools’ interfaces. Availability, scope and access conditions are checked during scoping.
Shared folders, procedures, office files and document repositories: organise useful sources to find information and connect it to its origin.
Documents, internal repositories and workspaces.
To assess: formats, versions, read permissions and update frequency.
Explore the solutionConnect forms, incoming enquiries and CRM records to prepare follow-up, qualify requests and avoid re-entering data.
CRM, email, forms and sales follow-up.
To assess: available APIs, reference fields and actions requiring approval.
Explore the solutionMove useful information between business tools: statuses, orders, cases or stock, with tracking of exchanges and errors.
ERP, business management and internal applications.
To assess: the authoritative source, identifiers, exports and synchronisation rules.
Explore the solutionIntegrate a model into a specific workflow: search, extraction, assisted writing or content creation. Results are presented in your tools.
Models through APIs and the Moon AI platform.
To assess: expected quality, transmitted data, usage costs and review steps.
Explore the solutionTrigger processing from an event, run a sequence of steps and make each operation’s status understandable to the team.
APIs, webhooks, scheduled processing and n8n.
To assess: error recovery, timing, write permissions and necessary notifications.
Explore the solutionPrepare a structured data foundation for applications, organise permissions and define how exchanges are monitored and maintained.
PostgreSQL, Supabase, internal services and infrastructure.
To assess: retention, backups, environments and account management.
Explore the solutionMoon AI is our product. Moon6 is published on Ollama. Their uses, models and conditions are considered alongside your organisation’s needs.
Discover Moon AICRM, collaboration platforms, GitHub services, n8n automation or models from Mistral AI: we assess available interfaces and the selected service’s conditions.
A tool’s presence in this ecosystem does not make it a studio partner. Proprietary services and source-available software retain their own terms of use.
Identify tools, users, data and their owners.
Test a limited workflow with explicit rules and suitable data.
Document flows, make errors visible and define who responds to incidents.
Open code makes auditing, adaptation and handover easier. Control also depends on hosting location, operators, administrative access, dependencies and data flows. We examine these dimensions together.
Yes, depending on the selected models, their licences and available resources. We assess expected quality, hardware, response times and maintenance conditions before choosing a local deployment or an API.
Responsibilities are defined in the project scope: updates, backups, recovery, monitoring and access. Documentation and handover arrangements are among the points to establish.
Integrate AI, build an application or rethink your data architecture. Tell us about your situation and we’ll discuss the options.
PhoneMonday to Friday, 10:00 to 22:00
Saturday and Sunday: email only
Paris time (Europe/Paris)