Felan
Cloud platform for running persistent AI agents (developer, QA, SRE personas) across a team's whole software development lifecycle, with shared memory and full session visibility.
🔗 Visit Felan
Description
Most AI coding assistants live inside one person's editor and forget everything the moment the session ends — which means a whole engineering team using AI agents ends up with dozens of agents that never learn from each other. Felan tries to fix that by running AI agents in the cloud instead, as background workers with different jobs (writing code, testing it, watching production) that share what they've learned across the whole team, not just one developer's machine.
Felan runs cloud-based background agents with defined personas (Developer, QA, SRE, General Purpose), triggered by events or schedules rather than only manual prompts, and it keeps a persistent, team-wide knowledge base plus full session transcripts so any teammate can see exactly what an agent did and why. It connects to 17+ existing tools (GitHub, GitLab, Slack, Teams, Linear, Jira) and structures work through predefined SDLC stages — Capture, Groom, Plan, Implement, Review — with bring-your-own-API-key model portability rather than locking a team into one AI provider. Felan is the direct successor to Bugzy, a QA-test-generation tool that merged into it; existing Bugzy customers are now Felan customers.
💬 Our review
The short version: Felan is betting that AI agents are more useful as persistent, team-shared cloud workers than as one-off assistants inside a single developer's editor, and it's building the shared-memory and multi-persona layer to make that real.
The genuinely different idea here versus Claude Code, Cursor, or GitHub Copilot running locally is persistence and team visibility — an agent's context and learned knowledge staying available to the whole team (not resetting per session, per developer) is a real gap in today's AI coding tools, and full session transcripts address the legitimate worry of "what did the agent actually do while I wasn't watching." The honest caveat: this is an early-stage product (it absorbed Bugzy's QA-focused customer base rather than growing this platform from scratch) still in a waitlist/early-access posture with no public pricing, so a team evaluating it should expect rough edges and a smaller integration ecosystem than the established AI coding agents it's implicitly competing against. If your team's actual need is specifically automated QA test generation — what Bugzy used to do — evaluate TestSprite, QA Wolf or Checksum directly rather than assuming Felan still serves that narrower need as well as its predecessor did.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Not publicly available; early access/waitlist stage with custom pricing implied.
Pros
Persistent, team-shared agent memory instead of per-session/per-developer context
Multiple agent personas (Developer, QA, SRE) for different SDLC stages
Full session visibility/transcripts for team-wide auditability
Bring-your-own-keys model portability, not locked to one AI provider
Cons
Early-stage, waitlist access, no public pricing yet
Absorbed a QA-specific predecessor (Bugzy) — no longer a QA-only specialist tool
Smaller integration ecosystem than established local coding agents