Roadmap and Intake
General Augment does not publish a full public roadmap today. Integration partners and customers get quarterly roadmap summaries and active launch lists in their working channel.
Signed Enterprise terms own any formal SLA. This page is the current product posture for teams deciding whether to build on General Augment.
Q2 2026 Direction
Section titled “Q2 2026 Direction”The current platform work is focused on making General Augment usable as the agent backend for multiple apps:
/v1/responsesas the stable reasoning API- per-user memory APIs and profile visibility
- governed tools, approvals, credential isolation, and audit logs
- response IDs, trace IDs, usage, token counts, and cost metadata
- dashboard, CLI, SDK, examples, and local mock testing
- channels and background jobs after the core API path is healthy
- security, reliability, deletion, quota, and budget guardrails
Reference Customer Policy
Section titled “Reference Customer Policy”Spark is the reference customer path for the current readiness cycle. Spark should not receive a private fork or special runtime. When Spark exposes a platform gap, the default is to turn that fix into a reusable General Augment capability for future apps.
Spark-specific assumptions should stay in Spark’s app backend or be documented as temporary until they become generally useful platform behavior.
Intake And Capacity
Section titled “Intake And Capacity”General Augment may limit or pause new third-party app onboarding when the platform cannot safely support the requested launch. Common reasons include provider quota, infrastructure capacity, compliance requirements, support load, unresolved security review findings, billing/readiness gaps, or incident recovery.
Active production or launch-path partners should have visible launch items, named owners for launch-critical gaps, and a recurring partner sync while the integration is moving toward production. Formal availability, response-time, compliance, or capacity commitments belong in signed customer terms.
High-volume apps should request a traffic-shape review before launch. Share expected request rate, burst windows, model tier mix, memory usage, tool usage, channel sends, retry behavior, and budget guardrails. Do not build against unpublished capacity numbers. This intake page is not an SLA; formal capacity, response-time, and availability commitments belong in signed terms.