Releases: cchinchilla-dev/agentloom
Releases · cchinchilla-dev/agentloom
v0.4.0
What's Changed
Internal
- trigger release workflow from auto-tag via workflow_run by @cchinchilla-dev in #93
- allow workflow_dispatch in release workflow condition by @cchinchilla-dev in #94
- fall back to nearest tag when HEAD has no tag by @cchinchilla-dev in #95
Other Changes
- add pluggable checkpoint backends and resume CLI (#78) by @cchinchilla-dev in #96
- add workflow pause mechanism for human-in-the-loop (#40) by @cchinchilla-dev in #97
- add approval gate step type for human-in-the-loop (#41) by @cchinchilla-dev in #98
- add webhook notifications for approval gates (#42) by @cchinchilla-dev in #99
- docs: add attribution doc by @cchinchilla-dev in #100
- refresh claude skills to match real workflow by @cchinchilla-dev in #101
- clean up decorative comments and align repo conventions by @cchinchilla-dev in #103
- add mock and recording providers for deterministic replay by @cchinchilla-dev in #102
- complete mock provider cli integration (#61, #76) by @cchinchilla-dev in #113
Full Changelog: v0.3.0...v0.4.0
v0.3.0
What's Changed
Documentation
- add mkdocs-material documentation site (#72) by @cchinchilla-dev in #91
Internal
- upgrade actions to node 24 and improve pr labeler by @cchinchilla-dev in #37
- add Ollama e2e integration tests and CI workflow by @cchinchilla-dev in #82
- add coverage reporting via Codecov and 85% threshold by @cchinchilla-dev in #84
Other Changes
- add YAML pricing config and AGENTLOOM_ env var support (#5, #6) by @cchinchilla-dev in #79
- add multi-modal input support (images, PDFs, audio) (#68) by @cchinchilla-dev in #80
- add streaming support for LLM responses by @cchinchilla-dev in #81
- add array index support in state paths by @cchinchilla-dev in #83
- expose DAG as first-class graph API by @cchinchilla-dev in #85
- cache gateway lookup and async checkpoint I/O by @cchinchilla-dev in #86
- emit budget_remaining metric to OTel by @cchinchilla-dev in #87
- add Grafana panel for budget_remaining metric by @cchinchilla-dev in #89
- add provider comparison table to README (#73) by @cchinchilla-dev in #90
- bump version to 0.3.0 by @cchinchilla-dev in #92
Full Changelog: v0.2.0...v0.3.0
What's Changed
Documentation
- add mkdocs-material documentation site (#72) by @cchinchilla-dev in #91
Internal
- upgrade actions to node 24 and improve pr labeler by @cchinchilla-dev in #37
- add Ollama e2e integration tests and CI workflow by @cchinchilla-dev in #82
- add coverage reporting via Codecov and 85% threshold by @cchinchilla-dev in #84
Other Changes
- add YAML pricing config and AGENTLOOM_ env var support (#5, #6) by @cchinchilla-dev in #79
- add multi-modal input support (images, PDFs, audio) (#68) by @cchinchilla-dev in #80
- add streaming support for LLM responses by @cchinchilla-dev in #81
- add array index support in state paths by @cchinchilla-dev in #83
- expose DAG as first-class graph API by @cchinchilla-dev in #85
- cache gateway lookup and async checkpoint I/O by @cchinchilla-dev in #86
- emit budget_remaining metric to OTel by @cchinchilla-dev in #87
- add Grafana panel for budget_remaining metric by @cchinchilla-dev in #89
- add provider comparison table to README (#73) by @cchinchilla-dev in #90
- bump version to 0.3.0 by @cchinchilla-dev in #92
Full Changelog: v0.2.0...v0.3.0
v0.2.0
What's Changed
Internal
- Add Kubernetes, Helm, Terraform, and ArgoCD deployment stack by @cchinchilla-dev in #35
Other Changes
- add Docker infrastructure and improve observability stack by @cchinchilla-dev in #34
- bump version to 0.2.0 by @cchinchilla-dev in #36
Full Changelog: v0.1.2...v0.2.0
v0.1.2
What's Changed
Bug Fixes
- fix sync state access in async step executors by @cchinchilla-dev in #30
Documentation
- update logos and header images by @cchinchilla-dev in #19
- update dark header image by @cchinchilla-dev in #20
Other Changes
- Sandbox enforcement for built-in tools (#4) by @cchinchilla-dev in #31
- bump version to 0.1.2 by @cchinchilla-dev in #32
Full Changelog: v0.1.1...v0.1.2
v0.1.1
What's Changed
Documentation
- add copilot code review instructions by @cchinchilla-dev in #1
- use absolute urls for readme header image by @cchinchilla-dev in #2
Other Changes
- fix rate limiter to account for response tokens by @cchinchilla-dev in #17
- release v0.1.1 by @cchinchilla-dev in #18
Full Changelog: v0.1.0...v0.1.1
v0.1.0
First public release
Production-ready agentic workflow orchestrator with native observability, resilience, and cost control.
Highlights
- YAML and Python DSL workflow definitions — DAGs with sequential + parallel steps
- Step types:
llm_call,tool,router(conditional),subworkflow - Provider gateway with automatic fallback (OpenAI, Anthropic, Google, Ollama)
- Resilience — circuit breaker, rate limiter, and retry with exponential backoff per provider
- Budget enforcement — hard stop when USD limit exceeded
- Cost tracking per step, model, and provider
- Observability — OpenTelemetry traces + Prometheus metrics (optional,
pip install agentloom[all]) - CLI —
run,validate,visualize(ASCII + Mermaid),info - Checkpointing — save and resume workflow state to disk
Design decisions
- httpx over provider SDKs — minimal dependencies (~5 core)
- anyio over raw asyncio — structured concurrency via task groups
- str.format_map over Jinja2 — one fewer dependency for prompt templates
- Observability optional — NoopSpan/NoopTracer pattern, zero overhead when not installed
- Pydantic v2 — validation and serialization for all models
Stats
- 392 tests, mypy strict, ruff clean
- 16 workflow examples included
Full Changelog: https://github.com/cchinchilla-dev/agentloom/commits/v0.1.0