Predictable prices, capped by your budget

Allow and Block Lists

Configuration of email and domain allow lists and block lists for sign-up control, restricting which addresses can create accounts in an auth app.

Restricted Signup Mode

Ability to restrict sign-ups to invitation-only or pre-approved accounts, preventing open registration on an auth app.

Custom Attributes

Definition of custom attribute schemas on end-user accounts for application-specific data that extends the default profile fields.

Enterprise SSO (SAML)

SAML-based enterprise SSO connections with assertion encryption and request signing for corporate identity provider integrations.

Enterprise SSO (OIDC)

Enterprise OIDC SSO connections for corporate identity providers that support the OpenID Connect protocol.

Enterprise SSO (Custom)

Custom enterprise SSO integrations for proprietary identity systems that do not conform to standard SAML or OIDC protocols.

JIT Provisioning

Just-in-time user provisioning that creates accounts automatically on first SSO sign-in without requiring pre-registration.

Domain Routing

Automatic redirect of users to the correct SSO provider based on their email domain during the authentication flow.

Enterprise Role Mapping

Automatic role assignment from external directory attributes during SSO authentication, syncing group memberships to Conjoin roles.

SCIM Provisioning

SCIM directory provisioning for automated user and group lifecycle management from external identity providers like Okta, Azure AD, or OneLogin.

Maximum Enterprise SSO Connections

Your plan includes a number of enterprise SSO identity-provider connections (SAML, OIDC, and custom) per auth app. Extra connections are billed per connection on metered tiers.

Maximum Social Login Connections

Maximum number of social login provider connections (Google, GitHub, Apple, and other consumer-IDP options) enabled per auth app.

MFA (TOTP)

Time-based one-time password (TOTP) authentication using authenticator apps like Google Authenticator or Authy as a second factor.

MFA (SMS)

SMS-based second-factor verification codes delivered to registered phone numbers during step-up authentication.

MFA (Passkey)

WebAuthn passkey authentication using hardware or platform authenticators as a second factor for phishing-resistant verification.

Passkeys as Primary Auth

Passkeys as a primary authentication method, enabling passwordless sign-in with WebAuthn credentials on supported devices.

Magic Links

Passwordless magic link authentication delivered via email, allowing one-click sign-in without password entry.

Rate Detection Suite

Auth Guard ships rate-based detectors that flag brute force, signup abuse, verification flooding, and dormant-account reactivation, so threat policies can act on rate-based attack patterns.

Bot Detection Suite

Auth Guard ships bot detectors that read signature and device-fingerprint signals, so threat policies can act on automated-actor activity.

Geo Detection Suite

Auth Guard ships geographic detectors that flag impossible-travel velocity, new-country sign-ins, and region restriction rules, so threat policies can act on location-based anomalies.

Network Detection Suite

Auth Guard ships network-source detectors that flag anonymous-network indicators (Tor, public VPN, residential proxy), so threat policies can act on network-origin signals.

Session Detection Suite

Auth Guard ships session-state detectors that flag concurrent-session limits and anomalies, so threat policies can act on session-lifecycle signals.

Email Detection Suite

Auth Guard ships email detectors that flag disposable email addresses from a managed domain set plus your own blocked domains, so threat policies can act on throwaway-identity signups.

Advanced Guard Rules

On Enterprise, you can author custom Auth Guard rules that compose signals from the rate, bot, geo, network, and session detectors with your own predicates and actions.

Session Device Binding

Cryptographic binding of sessions to device fingerprints, preventing session transfer between devices and strengthening token security.

Advanced Session Controls

Configure your auth app session policy beyond the platform defaults, including session and token durations, concurrent-session limits per account, and eviction behaviour when limits are reached.

Monthly Active Accounts

Total number of distinct end-user accounts that authenticate at least once per billing cycle, counted once per account regardless of how many sign-ins occur.

Monthly Active Organisations

Total number of distinct organisations counted as active per billing cycle. An organisation is active in a cycle when it has two or more members and at least one of them was an active account that cycle; single-member organisations are never counted.

Max Organisation Members

Maximum number of members allowed per auth organisation, controlling tenant size for capacity planning and plan-based limits.

Max Organisation Groups

Maximum number of groups allowed per auth organisation, controlling the complexity of group-based role assignment and SCIM sync.

Maximum Organisations

Your plan caps how many active organisations each auth app can hold, where every organisation acts as a tenant within the identity app.

Maximum Auth Apps

Your plan caps how many active auth apps you can provision per cloud account, where each app is an independent identity tenant.

Auth Data Retention

Maximum number of days auth records (sessions, trusted devices, invitations) are retained after expiry or status transition, controlling how long historical auth data is kept per plan.

Outbound Emails Per Month

Total number of outbound email messages sent per billing cycle, counted once per send regardless of how many recipients the message addresses. The maximum addresses per single send is bounded by the messaging:email:max-recipients-per-send entitlement.

Maximum Sending Domains

Your plan caps how many verified custom sending domains each project can configure for outbound email, with DKIM, SPF, and DMARC handled for every domain.

Maximum Email Recipients Per Send

Maximum number of recipient addresses (To, Cc, and Bcc combined) allowed on a single outbound email send. Caps fan-out per API call so bulk delivery is batched into multiple sends rather than a single jumbo request.

Email Open Tracking

Pixel-based open event tracking on outbound emails, recording when recipients view messages for engagement analytics.

Email Click Tracking

Link-rewriting click event tracking on outbound emails, recording which links recipients interact with for conversion analytics.

Email Unsubscribe Tracking

One-click unsubscribe tracking with automatic suppression list management, ensuring compliance with email sending regulations.

Maximum Email Senders

Maximum number of verified email sender identities (domain and address pairs) configured per project for outbound email delivery.

Maximum Email Sender Pools

Maximum number of email sender pools a project may define, where each pool groups verified senders for load distribution and failover.

Inbound Email Messages Per Month

Total number of inbound email messages received through configured inbound routes per billing cycle, counting each delivered message once.

Custom Email Tracking Domain

Availability of a customer-owned domain for open and click tracking links on outbound email, replacing the default Conjoin tracking domain.

Maximum Messaging Templates

Your plan caps how many stored messaging templates each project can hold across every channel, where each template is a reusable message body you can send against.

Maximum SMS Recipients Per Send

Maximum number of phone numbers allowed on the recipient list of a single outbound SMS. Caps fan-out per API call so bulk delivery is batched into multiple sends rather than a single jumbo request.

Outbound SMS Per Month

Total number of outbound SMS sends per billing cycle, counted once per send regardless of how many phone numbers the send targets. The maximum phone numbers per single send is bounded by the messaging:sms:max-recipients-per-send entitlement.

SMS Brand Registration

Access to 10DLC brand and campaign registration, required for high-throughput A2P SMS delivery in North America.

Maximum Registered SMS Brands

Maximum number of business identities registered with The Campaign Registry for 10DLC compliance and trust scoring per project.

Maximum Registered SMS Campaigns

Maximum number of A2P use-case campaigns registered against TCR brands for regulatory approval of high-volume SMS sending.

Maximum SMS Senders

Maximum number of registered SMS sender identities (originating numbers or alphanumeric sender IDs) configured per project for outbound SMS delivery.

Maximum SMS Sender Pools

Maximum number of SMS sender pools a project may define, where each pool groups verified senders for load distribution and failover.

Inbound SMS Segments Per Month

Total number of inbound SMS segments received on rented phone numbers per billing cycle, counting each carrier-delivered segment once.

Scheduled Sending

Queue messages for future delivery at a specified date and time, supporting timezone-aware scheduling for optimal engagement.

Suppression Lists

Automatic email suppression list management based on bounces, complaints, and delivery failures to protect sender reputation.

Verifications Per Month

Total number of verification attempts initiated per billing cycle across OTP flows on every supported channel.

Maximum Attempts Per Verification

Maximum number of code submission attempts allowed against a single verification before it is locked out and a new verification must be requested.

Verification Email Channel

Availability of email as a delivery channel for OTP verifications dispatched through the verification pipeline.

Verification SMS Channel

Availability of SMS as a delivery channel for OTP verifications dispatched through the verification pipeline.

Maximum Active Conversations

Maximum number of conversations that may be in the active (non-archived) state at any one time, capping concurrent two-way messaging threads.

Maximum Participants Per Conversation

Maximum number of recipients that may be attached to a single conversation, capping the participant fan-out for any one thread.

Maximum Conversation Recipients Per Project

Maximum number of conversation recipient records that may be stored across every conversation in a project, capping the total participant footprint a project can hold.

Conversation Retention Days

Maximum number of days conversation history is retained before messages and attachments are purged from project storage.

Maximum Messaging Profiles

Maximum number of messaging profile configurations a project may maintain, where each profile defines channel features, sender defaults, and tracking policies.

Maximum Contacts

Maximum number of recipient records stored in the messaging per project.

Maximum Phone Numbers

Maximum number of telephone numbers a project may rent from carriers for SMS across all regions.

Maximum Customers

Your plan caps how many active billing customer records you can keep per billing entity.

Maximum Active Subscriptions

Your plan caps how many active subscription records you can hold per billing entity at any one time.

Subscription Schedules

Multi-phase subscription schedules that swap plans, prices, and quantities on predefined future dates, enabling planned upgrades, downgrades, and trial-to-paid transitions without manual intervention.

Invoices Per Month

Your plan caps how many invoices each billing entity can issue per cycle, with overage priced per invoice on metered tiers.

Invoice PDF

PDF generation and download for customer invoices and receipts with customisable templates and branding.

Maximum Products

Your plan caps how many active product definitions you can keep in the billing catalog per entity, covering products together with their price configurations.

Credit Notes

Issue credit notes against finalised invoices for refunds, adjustments, and corrections. Every credit note links back to the original invoice for audit.

Credit Wallets

Each customer can hold a credit wallet with prepaid balances, promotional credits, and refund credits, applied automatically to future invoices.

Configurable Dunning

Configurable dunning workflows for failed-payment recovery, including retry strategy selection (aggressive, moderate, conservative), escalation rules, and customer-notification cadence. The basic three-retry preset ships on every plan; this entitlement unlocks configuration of the dunning workflow past that baseline.

Tax Calculation

Automatic tax calculation on invoices and payments using the Conjoin-managed jurisdiction catalog. Supports compound and cascading components plus B2B exemption and reverse-charge rules. Jurisdictions are administered centrally by Conjoin; merchants consume them read-only. Calculation only: registration-threshold monitoring, filing, remittance, exemption-certificate validation, and tax-ID verification are out of scope and remain the merchant's responsibility.

Tax Calculations per Month

Conjoin counts each successful tax calculation on an invoice or payment intent once per cycle. Usage above the included quota bills per calculation.

Metered Pricing

Usage-based metered price models with configurable aggregation and reset intervals for pay-as-you-go billing.

Tiered Pricing

Volume and graduated tier pricing schemes on products, enabling quantity-based discounts and progressive rate structures.

Renewal Notifications

Automated subscription renewal reminder emails sent to customers before billing cycle ends to reduce involuntary churn.

Overdue Reminders

Automated overdue payment reminder emails with configurable schedules and escalation rules for outstanding invoices.

Queue Tasks Per Month

Total number of queue tasks dispatched per billing cycle, covering both immediate and scheduled task submissions across all queues.

Maximum Queues

Maximum number of active queues allowed per project, controlling how many independent task processing pipelines can run concurrently.

Dead Letter Queue

Routing of permanently failed tasks to a dead letter queue for inspection, replay, and failure analysis after all retry attempts are exhausted.

Queue Rate Limiting

Per-queue dispatch rate controls that throttle task delivery to downstream targets, preventing target overload during high-throughput periods.

Per-Queue Dispatch Rate (req/s)

Maximum number of task dispatches per second allowed on a single queue, setting the ceiling for rate-limited delivery throughput.

Per-Queue Maximum Concurrent Deliveries

Maximum number of in-flight deliveries allowed at the same time for a single queue, capping parallel task processing concurrency.

Maximum Schedule Horizon (days)

How far into the future a task can be scheduled for delayed delivery, measured in days from the time of submission.

Maximum Task Retention (days)

Number of days completed and failed task records are retained before automatic purge, controlling how long task history remains queryable.

Maximum Batch Enqueue Size

Maximum number of tasks allowed in a single batch enqueue request, controlling per-request throughput by plan tier.

Event Messages Per Month

Total number of event messages published per billing cycle across all event channels in the project.

Maximum Event Channels

Maximum number of active event channels allowed per project for pub/sub message routing and subscriber fan-out.

Event Replay

Ability to replay historical event messages to subscribers from a specific point in time for recovery and reprocessing scenarios.

Event Filtering

Server-side event filtering on channels, allowing subscribers to receive only messages matching specified attribute patterns.

Maximum Broadcast Channels

Maximum number of active broadcast channels allowed per project for realtime WebSocket message distribution.

Broadcast Connection Minutes

Total connection-minutes consumed per billing cycle, summing every minute each WebSocket connection stays open across all broadcast channels in the project.

Maximum Concurrent Broadcast Connections

Maximum number of WebSocket connections allowed open at the same time across all broadcast channels in the project, enforced as a per-account safety ceiling above the included connection-minute allowance.

Broadcast Messages Per Month

Total number of broadcast messages sent per billing cycle across all broadcast channels in the project.

Presence Tracking

Realtime presence awareness on broadcast channels, tracking which users are connected and surfacing join/leave events to subscribers.

Channel Authorization

Per-channel authorization checks on broadcast connections, restricting access based on user identity and application-defined permissions.

Maximum Broadcast History Size

Maximum number of historical messages retained per broadcast channel for catch-up delivery when clients reconnect.

Maximum Schedules

Maximum number of active schedule definitions allowed per project, covering both cron-based recurring and one-time scheduled executions.

Schedule Executions Per Month

Total number of schedule executions fired per billing cycle across all schedule definitions in the project.

Execution History Retention Days

Number of days scheduler execution records are retained for audit and debugging before automatic purge.

Scheduler Dead Letter Queue

Routing of permanently failed scheduler executions to a dead letter queue for inspection, replay, and failure analysis.

Maximum Scheduler Timeout (seconds)

Longest allowed execution timeout in seconds for a single scheduler invocation before the system marks it as timed out.

MemoryStore Shared Storage (MB)

Maximum provisioned storage in megabytes across all shared-backing MemoryStore databases in the project. Dedicated databases are sized by what was bought and do not count against this quota.

MemoryStore Commands Per Month

Total number of MemoryStore commands executed per billing cycle, covering reads, writes, and atomic operations.

Maximum Shared MemoryStore Databases

Maximum number of MemoryStore databases on the shared multi-tenant backing pool that the account may provision concurrently across all projects.

Maximum Dedicated MemoryStore Databases

Maximum number of single-tenant dedicated MemoryStore databases that the account may provision concurrently across all projects. Zero indicates dedicated databases are not available on the plan.

Default MemoryStore Replicas

Default high-availability replica count provisioned for shared MemoryStore databases, controlling baseline failover capacity without explicit per-database configuration.

Maximum MemoryStore Replicas

Configurable upper bound on high-availability replica count per MemoryStore database, capping how many synchronous replicas a customer may attach to a single database.

TCP Access on Dedicated Databases

Direct TCP/RESP socket access to dedicated MemoryStore databases via rediss:// with TLS and IAM_AUTH; provisions per-customer infrastructure when enabled on a dedicated database. Shared databases include TCP/RESP access by default via the multi-tenant proxy fleet and do not require this entitlement.

Pipeline Runs Per Month

Total number of pipeline executions started per billing cycle across all pipeline definitions in the project.

Maximum Steps Per Run

Maximum number of steps allowed in a single pipeline run, capping workflow complexity for resource management.

Maximum Active Pipelines

Maximum number of active pipeline definitions allowed per project, capping how many independent workflow templates can be scheduled concurrently.

Maximum Pipeline Concurrency

Maximum number of pipeline runs that may execute in parallel across the project, controlling worker contention and downstream resource pressure.

Pipeline Execution Log Retention (days)

Number of days pipeline execution logs are retained for audit and debugging before automatic purge.

Parallel Execution (Fork/Race)

Fork and race step types in pipelines, enabling concurrent branch execution and first-to-complete selection within a single run.

Gate Steps

Gate steps in pipelines that pause execution until an external approval signal is received, enabling human-in-the-loop workflows.

Search Indexes

Maximum number of full-text search indexes allowed per project for structured document indexing and query.

Documents Per Index

Maximum number of documents that can be stored in a single search index, controlling index size for query performance.

Search Queries Per Month

Total number of search queries executed per billing cycle across all search indexes in the project.

Search Batch Indexing

Bulk-ingest documents into search indexes through batch APIs, so first-time population and large reindex operations run at higher throughput than per-document indexing.

Relay Hook Fires Per Minute

Maximum number of per-task hook fires (webhook callbacks, event broadcasts, queue chains, pipeline continuations) aggregated across every hook type per project per minute, so one task cannot saturate downstream targets.

Relay Callback In-Flight Cap

Maximum number of concurrent in-flight callback delivery tasks per project, preventing a single project from saturating the shared callback delivery queue.

Relay Data Retention

Maximum number of days terminal relay records (completed tasks, delivered events, expired schedules) are retained before automatic purge, controlling how long historical relay data is kept per plan.

Inference Requests Per Minute

Maximum number of inference API requests allowed per minute per project, acting as a rate limit ceiling to prevent burst overload.

Inference Input Tokens Per Minute

Maximum input-token throughput per minute per project. Pairs with the requests-per-minute cap so a tenant cannot drain the inference budget by sending jumbo prompts at the full request rate.

Inference Concurrent Requests

Maximum number of inference requests that can execute simultaneously per project, controlling concurrency to protect downstream model providers.

Multi-Model Inference

Routing inference requests across multiple model providers and model versions within a single project configuration.

Best-of-N Inference

Generation of multiple candidate completions per request with automatic selection of the highest-scoring result based on configured criteria.

Inference Synthesis

Aggregation of responses from multiple model providers into a single synthesised output for improved accuracy and reliability.

Batch Inference

Submission of bulk inference requests for asynchronous processing at reduced cost, suited for non-latency-sensitive workloads.

Included AI Inference Spend

Each plan ships with an AI inference allowance per billing cycle. Conjoin tracks accumulated model costs against the allowance and resets the counter at the start of each cycle.

Policy Checks Per Month

Total number of content policy evaluation checks executed per billing cycle across all inference and context requests.

Custom Policy Patterns

Maximum number of custom regex or keyword patterns that can be defined for content filtering and PII detection rules.

Custom Policy Topics

Maximum number of custom topic classifiers that can be configured for semantic content categorisation and blocking.

Topic Blocking

Blocking of inference requests or responses that match configured topic classifiers, preventing prohibited content from reaching end users.

Response Inspection

Post-generation inspection of model responses against policy rules before delivery, enabling output-side content filtering.

Log-Only Mode

Policy evaluation in observation mode where violations are logged without blocking requests, used for rule tuning before enforcement.

Policy Audit Export

Export of policy evaluation logs and violation reports for external compliance review and archival systems.

Bring Your Own Key

Registration of customer-owned API keys for third-party model providers, allowing direct billing and credential management outside Conjoin.

BYOK Custom Endpoint

Configuration of custom base URLs for OpenAI-compatible and self-hosted model endpoints when using bring-your-own-key credentials.

Context Indexes

Maximum number of vector context indexes allowed per project for document embedding, retrieval, and RAG workflows.

Context Queries Per Month

Total number of context retrieval queries executed per billing cycle across all context indexes in the project.

Context Indexed Storage (bytes)

Maximum total storage in bytes for embedded document content across all context indexes in the project.

Hybrid Search

Combined vector similarity and keyword search on context indexes, merging semantic and lexical ranking for improved retrieval accuracy.

Agentic Context Access

Programmatic context index access from agentic workflows, allowing autonomous agents to query and retrieve documents during execution.

Context Query Reranking

Cross-encoder reranking pass on context retrieval results, reordering candidates by relevance before injection into the model prompt.

Per-Request Inference Traces

Per-request trace records capturing prompt, completion, model, provider, latency, and policy outcomes for every inference call, surfaced in the observability dashboard for debugging and audit.

Inference Latency Breakdowns

Latency breakdown per inference request covering routing, policy evaluation, model call, and post-processing phases, exposed in dashboards and trace records.

Budget Threshold Webhooks

Outbound webhooks triggered when AI spend crosses configurable budget thresholds within a billing cycle, enabling customer-side spend alerts and circuit-breaker integrations.

Usage Export

Export of AI usage records (inference traces, token counts, costs, policy outcomes) to external systems for SIEM ingestion, data warehousing, and finance reconciliation.

Observability Retention (days)

Number of days AI observability records (traces, latency breakdowns, usage exports) are retained before automatic purge, separate from compliance-side inference log retention.

AI Data Retention

Maximum number of days AI records (inference logs, policy evaluations, context queries) are retained after completion, controlling how long historical AI data is kept per plan.

Full Workflow Step Library

Access to the full agentic workflow step library, including advanced control-flow steps (fork, race, gate). Lower-tier plans receive a configured subset selected by the workflow engine; this entitlement unlocks the complete authoring surface.

Resident Storage (MB)

Conjoin tracks the time-weighted average of bytes you hold in storage across every container per billing cycle. Deletes decrement the counter, so the meter reflects what is resident rather than the total ever uploaded.

Maximum File Size (bytes)

Each plan sets a hard upload-time ceiling in bytes for a single object. Conjoin enforces the ceiling on direct uploads and on multipart commit.

Maximum Containers

Your plan caps how many active containers (buckets) each project can hold.

Object Versioning

Object versioning on containers, retaining immutable snapshots when objects are overwritten for rollback and audit purposes.

Lifecycle Automation

Automatic object deletion and storage class transition policies on containers based on age, access frequency, or custom rules.

Storage Class Transition

Automatic movement of objects between storage classes based on age rules, optimising cost by shifting infrequently accessed data to cheaper tiers.

Automatic Class Selection

Conjoin keeps each object in the cheapest storage class its read rate justifies, measuring how often the object is read rather than how old it is, because retrieving an object from a colder class carries a charge that an age rule cannot see. Turn it off on a container to pin a class and pay that class rate.

Delivery Bandwidth (MB Per Month)

Total bytes delivered out of storage through the CDN and direct download paths each billing cycle, summed across every container. Overage above the included quota is priced per byte.

Image Transformations Per Month

Conjoin counts every on-the-fly image transformation request served from your containers in the cycle. Usage above the included quota is priced per transformation on metered tiers.

Image Optimization

Access to the on-the-fly image transformation pipeline for resize, format conversion, and basic adjustments served through the CDN edge.

Advanced Image Optimization

Advanced image operations including watermarking, face detection, face blur, colour adjustments, and still frame extraction from video.

Conjoin Auth ACL

Integration with Conjoin Auth for object-level and container-level access control, enabling token-scoped permissions on storage resources.

Available Regions

Number of geographic regions available for container placement, controlling data residency and latency optimisation options.

Available Storage Classes

Number of storage class tiers available (hot, warm, cool, deep) for cost optimisation based on access patterns and retention needs.

Custom Domains

Mapping of customer-owned domains onto storage delivery endpoints for branded asset hosting, with TLS certificate management and CDN edge binding.

Maximum Shared Sirannon Databases

Your plan caps how many SQLite databases you keep on Sirannon shared servers per project. A shared server holds the databases of many customers and sends every write to its replicas on other machines. Databases on a dedicated cell of your own stay outside this cap, because the capacity of the cell limits them.

Maximum Dedicated Sirannon Cells

How many dedicated Sirannon cells you may hold per project. A cell holds three Sirannon servers on three engine nodes in separate failure domains, and you are the only customer with databases on it, so it holds reserved capacity for every hour that it exists and bills by its size.

Rows Read Per Month

Every row that your statements read on a shared Sirannon server during the cycle, counted as the rows that a scan passes over and not the rows that it returns, so a scan of 5,000 rows that returns five counts as 5,000 and a lookup by primary key counts as one. Index entries, trigger work, and foreign-key actions all count, a statement that fails counts what it read before failing, and the reads that a live query makes count as well. Change delivery, device sync, backups, and the changes that one node applies from another count nothing. Reads inside a dedicated cell of your own count nothing either, because the hourly price of the cell already covers them.

Rows Written Per Month

Every row that your statements insert, update, or delete on a shared Sirannon server during the cycle, together with every index entry that a write inserts, so one insert into a table with five indexes counts as six rows. The primary counts each write once, including the writes that a replica forwards to it. Writes inside a dedicated cell of your own count nothing, because the hourly price of the cell already covers them.

Resident Database Storage (MB)

The bytes that you hold across every database on both engines, which Conjoin samples once an hour and averages over the cycle. When you delete rows, the next sample reads the smaller size. Every byte rests in three copies across separate failure domains, and the meter counts your single copy. Storage meters the same way on a shared server, on a dedicated cell, and on a Postgres cluster.

Maximum Postgres Clusters

How many Postgres clusters you may hold per project. CloudNativePG operates each cluster as one primary and two synchronous standbys on separate failure domains, so the cluster holds reserved capacity for every hour that it exists and bills by its size and not by the rows that it serves.

Maximum Read Replicas Per Cluster

How many read replicas you may add to a cluster beyond the two synchronous standbys that durability needs. Each one answers read queries of its own and bills at the rate of the size that you choose for it.

Maximum Concurrent Connections

How many connections you may hold open at once across both engines, counted per project. Postgres connections arrive through the connection pooler, and Sirannon answers its own over HTTP and WebSocket.

Maximum Credentials Per Database

How many sets of credentials you may issue against one database or one cluster, so that each of your applications holds its own and you can revoke one without disturbing the rest.

Backup Retention (Days)

How long Conjoin keeps your backups. Postgres archives its write-ahead log and its base backups to the object store of the site, and Sirannon snapshots each database on the same schedule.

Manual Backups

Take a backup on demand through the API or the dashboard, at any point between two scheduled runs.

Point-In-Time Recovery

Restore a database to any moment inside the retention window, which Postgres replays from its archived write-ahead log.

Change Streams

Subscribe to the changes that a database commits, which Conjoin Relay delivers like every other service event. Turning this on raises your read count on a multi-row update, because SQLite runs an update in two passes on a table that has a trigger, so each updated row reads once or twice more.

Multi-Region Databases

Place a database, a dedicated cell, or a read replica in a region other than the home region of your project, so that reads reach the people making them from nearby.