Updates
Changelog
A timeline of everything new, fixed, and improved on the Lobstack platform.
✅
Stay up to date
Follow @lobstackai on X for announcements, or check this page for detailed release notes.
Releases#
v0.20.02026-06-04
Added
- -Lobstack Pay (LSP-1) — agentic settlement protocol: policy-governed smart-account wallets, the x402 payment-required handshake, USDC on Base, and USD-pegged Lobstack Credits
- -On-chain settlement — the proprietary LobstackSettlement contract settles USDC and emits verifiable receipts; gasless ERC-4337 smart accounts (paymaster-sponsored) plus a custodial EOA path
- -Operator — agents that build and run a business: objectives → an LLM-generated plan → an autonomous execution loop (agents work each step with their tools) → live KPIs → per-goal spend budgets
- -Overseer — the fleet brain: observes privacy-safe metrics, learns (rule engine + LLM analyst on any model), proposes actions you approve in the admin panel, applies them live, and measures the outcome to learn
- -Org/tenant scoping for the Overseer — per-org rollups, org-tagged insights, and an admin org filter for enterprise isolation
- -10 native enterprise integrations — QuickBooks, Xero, Bill.com, Ramp, Okta, ServiceNow, Confluence, Box, Segment, and Snyk
- -Wallet and Operator dashboard tabs; an Overseer admin panel with approve / apply / dismiss
Improved
- -Agent bridge v20 — Lobstack Pay tools (check balance, pay, auto-x402), native enterprise integration handlers, and Operator tools (goals, steps, KPIs) with a fire-and-forget execution endpoint
- -The LLM analyst and Operator planner run on the Lobstack Gateway — pick any model on managed Lobstack AI
v0.18.02026-06-02
Added
- -Lobstack AI Gateway — a unified, OpenAI-compatible endpoint in front of every provider, now the primary path agents use to reach a model
- -Use Lobstack AI (managed) — deploy an agent with no API key; model usage is included in your plan
- -Auto model selection — Token Intelligence routes each message to the cheapest capable model, bounded by your plan tier
- -Message-credit top-ups — buy extra messages from the dashboard; they're used automatically once your monthly allowance runs out
Changed
- -Bring-your-own-key agents now also route through the gateway for Token Intelligence + unified usage analytics (a single-provider key is never used for another vendor)
- -Switch between Lobstack AI and your own key, or toggle Auto, from Settings with zero downtime
v0.17.02026-06-02
Added
- -Integrations Directory — the full catalog of 117+ integrations is now browsable in the docs, grouped by category and sourced directly from the live skill catalog
- -Agent Templates page now surfaces all 128 production-ready templates across 17 categories (previously only a subset was shown), each pre-loaded with a role, tools, and recommended skills
- -Self-optimization loop (v0) — agents safely improve their own system prompts behind a frozen held-out eval gate, with atomic, reversible promotions
Changed
- -The Lobstack Agent Runtime is now documented as the proprietary, primary execution engine on every agent VM — driving LLM reasoning, real tool execution, multi-model routing, and memory extraction
- -Consolidated all compute onto Hetzner Cloud; provider references across docs, marketing, and legal pages updated accordingly
- -Homepage and Features repositioned around deploying ready-made agents out of the box, rather than building agents from scratch
- -MCP servers are now documented as natively supported by the Lobstack runtime
v0.16.02026-03-05
Added
- -Token Conservation Engine — smart model routing based on complexity scoring (0-100) across 5 tiers (nano → flagship), with context window optimization and token budget management, saving 40-60% on API costs
- -Chat-to-Automation pipeline — create workflows directly from conversation using /automate, /schedule, /trigger, and /remind slash commands with natural language time parsing
- -Agentic Session Management — sessions now support categories (general, task, research, support, automation, creative, debug), statuses (active, completed, paused, archived), pinning, and grouping
- -10 new slash commands: /automate, /schedule, /trigger, /remind, /status, /pin, /category, /done, /export for chat-driven automation and session management
- -Cron trigger execution engine — every-minute cron evaluates all enabled workflow triggers with full cron expression parsing and double-fire prevention
- -Billing safety-net cron — hourly job catches agents whose billing period elapsed but wasn't reset by Stripe webhooks
- -Webhook retry with exponential backoff — 3 attempts at 0s, 5s, 30s with smart retry logic (skips 4xx except 429)
- -Enhanced analytics — per-model token breakdown, hourly aggregation for peak detection, and composite efficiency scoring
- -Token budget API — GET returns budget status and burn rate, POST returns model routing recommendations with savings estimates
- -/export command — downloads conversation as formatted markdown file
Fixed
- -Dashboard template dropdown — replaced overflowing horizontal category buttons with proper select dropdown in AgentConfigPanel
- -N+1 query in health cron — batch fetch all agent logs instead of per-agent loop queries
- -Admin listUsers() N+1 — replaced loading all Supabase Auth users with targeted email-based DB query in user sync and assign-user routes
- -Dashboard over-polling — tiered polling (15s fast for status/tokens, 60s slow for analytics/workflows) reduces API calls ~40%
Improved
- -Chat session sidebar — grouped by Pinned → Active → Completed → Archived with category icons, pin indicators, and context menu for rename/pin/categorize/complete
- -SELECT * elimination — admin routes now fetch only required columns for agents, users, and logs queries
- -Session API rewritten — GET/POST/PATCH/DELETE support category, status, icon, pinned, and message_count metadata
- -Slash command registry expanded to 22 built-in commands with rich descriptions and examples
v0.15.02026-03-03
Added
- -8 new AI model providers: DeepSeek (V3, R1), Qwen (Max, Plus), Mistral (Large, Small), Meta Llama 3.3 70B via Groq, Ollama self-hosted, expanding from 16 to 24 supported models
- -Edit workflow functionality — modify existing workflows (steps, triggers, name, description) without deleting and recreating
- -Redesigned Settings model selector with provider-grouped layout, color-coded provider headers, and model badges (Latest, Fast, Cheap, Free, Reasoning, Open Source)
- -Expanded bento box on landing page from 3 to 8 models with cycling carousel and color-coded category badges
- -SVG brand icons for Google, DeepSeek, Mistral, Meta, Groq, Qwen, and Ollama
- -Ollama self-hosted model support — runs locally on agent VM with no API key required
- -Firecrawl Scrape tool added to workflow builder with default markdown output format
Fixed
- -Critical workflow execution JSON parse error ("Extra data: line 1 column 2006") — _truncate_result was creating malformed JSON by truncating mid-stream and appending invalid suffix
- -Workflow variable interpolation double-serialization — dict/list values from step outputs were being stringified twice, breaking downstream tool inputs
- -Firecrawl Scrape field name mismatch — builder used "formats" (plural) but bridge expected "format" (singular), causing format parameter to be silently ignored
- -Agent Messages metric race condition — non-atomic read-then-write lost counts under concurrency; now uses ground-truth DB count
- -Messages counter backwards jumps — status endpoint could compute lower count than cached; fixed with Math.max(dbCount, cachedCount)
- -Performance Intelligence chart flickering blank every 10-20 seconds — polling reset loading state unnecessarily; now uses stale-while-revalidate pattern
- -Workflow execution no longer injects status messages into Chat — results shown only in Sandbox Executions tab where they belong
- -Incorrect social links in docs pages — Discord, Twitter/X links now match Footer
- -SettingsPanel API key provider detection for DeepSeek, Qwen, Mistral, Groq, and Ollama models
Improved
- -Settings panel model selection UX — flat 24-item radio list replaced with 10 provider groups, badges, and scrollable layout
- -Workflow builder UX — Edit button in workflow actions, builder shows "Save Changes" vs "Create" based on mode
- -API key section dynamically adapts to provider — shows correct key format placeholder and provider name for all 10 providers
- -Ollama models show "Self-hosted — no API key needed" instead of key input
- -Documentation updated: Models page expanded to 24 models with 10 providers, Workflows page includes Edit feature and best practices, Changelog updated
v0.14.02026-03-03
Added
- -Agent Templates expanded from 20 to 36 across all 9 categories
- -syncToBridge retry helper with exponential backoff for critical bridge sync points
- -Skills panel error banner with retry button when catalog fetch fails
- -Agent Templates bento card on landing page hero section
Fixed
- -Critical AGENT_SECRET → AGENT_API_SECRET env var typo in webhooks — silently broke all event-triggered workflow executions
- -TOCTOU race condition in pending execution pickup — atomic UPDATE+SELECT claim pattern prevents duplicate execution
- -Execution status validation — invalid statuses now rejected with 400 instead of silently defaulting to "completed"
- -Execution POST auth — X-Agent-Id header preferred over untrusted body agent_id
- -Skills POST/PUT missing DB upsert error handling — failures now return 500 instead of silent continue
- -Skills PUT not regenerating SKILLS.md — config saves now update system prompt and push to bridge
- -Triggers PATCH missing bridge sync — updates now forward to VM like POST and DELETE
- -Workflows PATCH missing triggers in bridge sync — updates now include trigger data
- -Webhooks route handlers (GET/POST/DELETE/PATCH) missing try-catch — all now have safe JSON parsing
- -Webhook delivery silent error swallowing — failures now logged with context
- -Event trigger pipeline — queue-first pattern ensures execution record exists before bridge call
- -SkillsPanel catalog fetch not checking response.ok — now detects HTTP errors with user-facing error state
- -SkillsPanel handleSaveConfig double-parsing and missing failure handling
Improved
- -Workflow execution pipeline hardened end-to-end: event trigger → queue → bridge pickup → result reporting
- -Bridge communication reliability with retry and backoff for transient failures
- -Skills integration lifecycle fully verified: toggle, configure, sync, and system prompt regeneration
v0.13.02026-03-03
Added
- -12 new Agent Templates: Legal & Compliance, HR & Recruiting, E-Commerce Manager, DevOps & SRE, Education & Tutoring, Financial Advisor, Project Manager, Cybersecurity Analyst, Creative Director, Real Estate Agent, Healthcare Assistant, Podcast Producer
- -3 new template categories: Operations, Finance, Education
- -Slash command highlighting in Chat — commands render as styled tags with hash icon
Fixed
- -Critical timestamp conversion bug in execution API — now handles ISO strings, unix seconds, and milliseconds
- -Race condition in pending execution pickup — disabled workflows are now cancelled instead of executed
- -Missing UUID validation on agent_id header in pending executions endpoint
- -Missing error handling on trigger deletion in workflows API
- -Missing try-catch on request body parsing in executions POST
- -Invalid CSS class wrap-break-word → break-words in SandboxPanel file viewer
- -Unused state variable in SandboxPanel template selection
- -Redundant cursor ternary logic in WorkflowCanvas
Improved
- -Execution status validation against enum whitelist (pending, running, completed, failed, cancelled)
- -Agent template library expanded from 8 to 20 templates covering 9 categories
- -Workflow execution API error handling and robustness across all endpoints
v0.12.02026-03-02
Added
- -Agent Template Library with 8 pre-built configurations (Customer Support, Research Assistant, Content Creator, Developer Assistant, Sales & Outreach, Personal Assistant, Data Analyst, Social Media Manager)
- -Template browser UI in Agent Config panel with category filtering and one-click apply
- -Architecture documentation page with system diagrams and data flow visualizations
- -This changelog page for tracking all platform updates
- -Error boundary component for dashboard crash resilience
Fixed
- -All ESLint errors and warnings across the entire codebase
- -Auth callback React anti-pattern (setState in useEffect)
- -Unused variables and imports cleaned up across API routes and docs pages
- -Sitemap now reflects actual documentation routes
Improved
- -Code quality: zero TypeScript errors, significantly reduced ESLint warnings
- -Documentation site navigation updated with Architecture and Changelog sections
v0.11.02026-03-01
Fixed
- -agent_configs table not visible in PostgREST schema cache — added v12 migration with NOTIFY pgrst reload
- -aiohttp breaking agent bridge entirely — added stdlib HTTP fallback for both server and client
- -Deploy crashes on schema cache miss — wrapped config seeding in try/catch
- -Infrastructure pip install fragility — split aiohttp from core deps, added build essentials
v0.10.02026-02-28
Added
- -Deep runtime integration with MCP servers, the skill registry, and security hardening
- -150+ tool integrations: AI providers, crypto, finance, DevOps, CRM, publishing
- -Slash command system with enhanced + modal and rich output rendering
- -Queue-based workflow execution engine with heartbeat-driven bridge pickup
- -Visual workflow builder with Run Now feedback
- -Master Terminal for admin panel operations
- -Video script for Lobstack product demo
Fixed
- -Critical aiohttp scoping bug breaking ALL tool integrations
- -Workflow Create crash and Run Now not working
- -Stuck pending executions — keep pending until bridge confirms
- -Shared getAgentForUser utility — eliminated lookup divergence causing 404
v0.9.02026-02-27
Added
- -Redesigned Skills panel with modern categorized filtering
- -GlassFlow-style bento grid with product UI mockups on landing page
- -Dashboard UX overhaul with glass effects, collapsible sidebar, Performance Intelligence chart
- -20+ new skill integrations with real brand logos
- -User sync endpoint for reconciling Stripe customers with Supabase
- -Admin: assign-user endpoint and demo account creation
Improved
- -Smooth sidebar animation with proper corner radius
- -Consistent curved border radius across all dashboard pages
v0.8.02026-02-27
Added
- -Comprehensive SEO & AISEO optimization for search engine and LLM discoverability
- -30+ page documentation site covering all platform features
- -Modern floating pill navbar for landing page
- -Supabase auth email templates with Lobstack branding
Fixed
- -Subscription loss after Privy → Supabase auth migration
- -Checkout 401 errors — use in-memory session instead of cookie-based getSession()
- -Bearer token fallback for all API calls
- -Magic link auth and PKCE race conditions
v0.7.02026-02-26
Added
- -Complete Framer-grade light-first design system
- -Supabase Auth (replacing Privy) with Google OAuth + email magic links
- -GitHub OAuth sign-in
- -3D Spline hero with mouse-tracking tilt effect
- -New Lobstack shield lobster logo
Changed
- -Auth system migrated from Privy to Supabase
- -All pages aligned to light design tokens