Product Roadmap
What we've built.
And what's coming next.
Podcode runs frontier-grade coding agents on open models, inside a cloud GPU account you control. Here's what's live in the engine today, and what's next.
Active
What's live today.
Unified Canvas
Workspaces hold projects. Sessions hold conversations. Everything in one place.
Plan & Approve
Inline plan approvals that write directly to your repo as markdown documentation.
Live HUD Telemetry
Context budget, silent-time warnings, and per-session spend caps.
RunPod Integration
Deep deployment table for Persistent Pods and Serverless endpoints with daily spend caps.
Local Inferencing
Ollama, LM Studio, vLLM, and llama.cpp - code that never leaves your machine.
Keys in Your OS Keychain
Provider keys held in the OS keyring, encrypted at rest - read locally, never synced to us.
Provider Freedom
Anthropic, OpenAI, RunPod, Ollama and more as equal citizens. Swap per task - no lock-in.
Diff-Aware Rollbacks
Inline changes panel shows every file touched with one-click rollback.
Agent Corporations
Researcher, Builder, Reviewer, and other agent profiles.
Command Palette
⌘K to navigate sessions, swap models, and manage workspace settings instantly.
Background Watchdog
Daemon keeps tabs on idle pods even when Podcode is closed.
The horizon
What's coming next.
Infrastructure & Providers
More GPU providers
Expanding native deployment tables to additional GPU hosts.
Auto Model Rotation
Automatically switch models when limits are reached or models fail.
Agentic Orchestration
AI agents act not as mere assistants, but as autonomous employees with distinct domains of ownership.
Security & Enterprise
Enterprise Compliance
SOC 2 Type II, GDPR, and HIPAA-ready auditing capabilities.
Team Scale
Shared agent profiles, memory, and deployment templates. SSO/SAML/SCIM for organizations.
Compliance Audit Trail
Exportable, retention-controlled per-action logging for SOC 2 / regulated environments.