Your databases and storage
Deliver to your database or storage through a suitable connection or handler. Prepare files directly on the agent.
Agents on your infrastructure
Collect data on your side.
Manage it in Akmelo.
Run collection, databases, and exports on your servers. Decide who receives results and how long to keep them. Collected records can stay entirely on your side; platform synchronization is optional.
Your data.
Your way of working.
Choose where collection runs, who has access, and where results go.
Execution on your servers
Your team chooses servers, resources, and recipient access. An administrator approves addresses and permissions. Your internal database does not need to be public.
Familiar management
Define sources, scenarios, and schedules in Akmelo. Schedules already synchronized continue during temporary platform outages.
Store it with you.
Exchange on your terms.
Collection, record history, and local exports run in your infrastructure. You can enable result transfer to the platform separately.
Akmelo
Sources, scenarios,
schedules, and control
Your infrastructure
Your agent
Collection and processing
Records and history
Files and exports
Your database and storage connections
Result synchronizationOptional
Copying collected records to Akmelo is a separate setting. Leave it off if results must stay in your environment. The arrows above distinguish job management from operational communication.
Stays with you
- Collected records and their change history
- Local files and export results
- Connections created in the agent interface
Connected to the platform
- Source settings, scenarios, and schedules
- Agent communication, statuses, and run counts
- Error messages and execution logs
Record mirroring to Akmelo is enabled separately. Do not put confidential values in scenario code, errors, or logs: those details may be available to the platform.
Your recipient.
Your delivery route.
Connect your database or keep exports on the agent server. You decide what data is sent and where.
Your database
Use a suitable connector or handler on your side. Write data directly or deliver it through a prepared export.
PostgreSQL
MySQL
MariaDB
MongoDB
Files on your server
Prepare CSV, JSON, and XML locally. Keep files on the agent or send them to your storage and application along an agreed route.
.csv.json.xmlS3
Access within your network
For a private recipient, use an approved network or an SSH tunnel configured by your administrator. There is no need to expose the database publicly.
AgentSSH / networkYour system
Agents for your tasks
Start with one.
Grow your network.
Use separate agents for a client, business line, or region. Add capacity as tasks grow and distribute sites among agents.
Your plan and any additional agents determine how many you can use.
View plan and limitsServers under your control
Agent 1Your project
Agent 2Your project
Agent NNew task
Deploy with Docker
Manage from AkmeloWhen it helps a business.
Four workflows where placement, recipients, and access should be agreed in advance.
Studios and client projects
Akmelo agents for studios: collect catalogs in the client's infrastructure, use local connections, separate results, and hand over the project clearly.
See the workflow
Purchasing and manufacturing
Supplier-offer collection with an Akmelo agent: prices, stock, and lead times in the company's database, with local access and a processing schedule.
See the workflow
Retail and service networks
Akmelo agents for networks: regular regional offer collection, your own storage, local history, and data prepared for analysis.
See the workflow
Research teams
Research on your own infrastructure with Akmelo: local datasets, databases and files, access controls, and separation of results from operational data.
See the workflowData placement and protection
Keep data in Russian infrastructure.
Place agents, databases, recipients, and backups in Russia. This can support the data localization requirements of Article 18(5) of Federal Law No. 152-FZ when the other processing requirements are also met.
Read Article 18Storage under your control
You determine placement, staff access, retention, and backup rules. Use your own infrastructure or a Russian provider.
Your organization sets processing rules
Alongside placement, consider processing grounds, security measures, and every transfer route, including logs and external services. The law provides exceptions to localization requirements; owning a physical server is not mandatory.
More control.
Your responsibility.
An agent makes sense when data requirements justify maintaining your own environment.
What you gain
- Choice of storage location and provider
- Local connections and recipient access
- Results beside your processing system
- Accepted schedules continue during a connection loss
What you maintain
- Server costs, resources, and availability
- Updates, backups, and recovery
- Network rules, HTTPS, and minimum privileges
- Review of collected data and operational logs
An agent remains connected to the platform and subscription terms. It is not a standalone copy of Akmelo. Source rules and request-rate limits still apply.
For a personal project
A personal server can help developers too.
A parts catalog, personal search, or a research tool: keep data beside your application and collect on a schedule.
Developer scenario
Before installation
How do I deploy an agent?
Prepare a server with Docker. Install the prepared stack manually or arrange SSH deployment. Manual installation does not require giving the platform SSH access.
What reaches the platform?
With mirroring off, records stay on the agent. Settings, statuses, counts, errors, and logs remain connected to the platform. Do not put secrets or record contents in code or logs. Enable result mirroring separately.
Does the server need inbound access?
For normal synchronization, the agent initiates outbound HTTPS connections. Configure access to its local interface and exports separately. SSH installation requires separately agreed access.
Start with one source.
Choose a server, assign collection to an agent, and verify the delivery route. Expand once results and storage rules are clear.