v1.2.15

Orla v1.2.15

Release v1.2.15 of Orla.

Changelog

• 45d822d Update README.md
• f85838b feat: added skills access control foundation
• e579e66 feat: separate access control api
• 553d94b feat: updated open api spec for access control

---

Full Changelog: pyorla-v1.0.7...v1.2.15

v1.2.14

Orla v1.2.14

Release v1.2.14 of Orla.

Changelog

• 59b8610 feat: modified access control precedence

---

Full Changelog: pyorla-v1.0.6...v1.2.14

v1.2.13

Orla v1.2.13

Release v1.2.13 of Orla.

Changelog

• 4cf3a46 Update README.md
• f2bebe0 feat: added access control to orla
• d1147d5 feat: added access control to orla

---

Full Changelog: pyorla-v1.0.5...v1.2.13

v1.2.12

Orla v1.2.12

Release v1.2.12 of Orla.

Changelog

• 8837a07 feat: added hotpotqa
• a8edad1 feat: completed gsm8k experiment
• 19da0f9 feat: made scheduling policies uniform
• 78be14d feat: separate stage and request priorities

---

Full Changelog: v1.2.11...v1.2.12

v1.2.11

Orla v1.2.11

Release v1.2.11 of Orla.

Changelog

• 57f3cd5 feat: added cost model
• 5708324 feat: added cost model
• e2be725 feat: added vllm example
• b75d749 feat: adding gsm8k benchmark
• f3d223e feat: better logging for vllm docker
• db3adae feat: updated example
• fb219fe feat: updated readme
• 6384220 feat: updated readme
• 999be01 feat: updated readme
• f92fe57 feat: updated readme
• f327f6a fix: metric serialization bug

---

Full Changelog: pyorla-v1.0.1...v1.2.11

v1.2.10

Orla v1.2.10

Release v1.2.10 of Orla.

Changelog

• b7babda fix: release

---

Full Changelog: v1.2.9...v1.2.10

v1.2.9

Orla v1.2.9

Release v1.2.9 of Orla.

Changelog

• e32834c feat: adding pyorla
• eef6ba1 feat: centralized mock provider and server
• 11bf011 feat: code hygiene and documentation
• 115a06c feat: configurable queue capacity
• d36b1f7 feat: improving testing and code quality
• 04c0b6a feat: moved to harvard-cns
• 179b929 feat: observability and production-readiness
• ac72901 feat: reliability hardening
• 4eb6ca0 feat: switched to orla debug logs
• 2ede8f8 fix: ci failures
• 7d8e05f fix: docker compose for vllm
• 93488a1 fix: docker compose for vllm
• b3437ea fix: removed binary
• 7dbbcd6 fix: tool call bug in langgraph pyorla

---

Full Changelog: v1.2.8...v1.2.9

v0.0.0-test-20260323-001720

Orla v0.0.0-test-20260323-001720

Release v0.0.0-test-20260323-001720 of Orla.

Changelog

• e32834c feat: adding pyorla
• eef6ba1 feat: centralized mock provider and server
• 11bf011 feat: code hygiene and documentation
• 115a06c feat: configurable queue capacity
• d36b1f7 feat: improving testing and code quality
• 04c0b6a feat: moved to harvard-cns
• 179b929 feat: observability and production-readiness
• ac72901 feat: reliability hardening
• 4eb6ca0 feat: switched to orla debug logs
• 2ede8f8 fix: ci failures
• 7d8e05f fix: docker compose for vllm
• 93488a1 fix: docker compose for vllm
• b3437ea fix: removed binary
• 7dbbcd6 fix: tool call bug in langgraph pyorla

---

Full Changelog: v1.2.8...v0.0.0-test-20260323-001720

v1.2.8

Orla v1.2.8

Release v1.2.8 of Orla.

Changelog

• d4f73d7 feat: SWE-bench Lite refactor - baseline package, Makefile, docker-compose.swebench-lite
• dfbfceb feat: add sb-cli project
• 052915e feat: added a healthcheck to sglang
• 452f646 feat: added better api for backends
• 69168ee feat: added better routing
• 5498b32 feat: added chat template kwargs
• ba061f3 feat: added dag math
• 95f4c44 feat: added dataset
• fc6c297 feat: added demo tape
• 4c10535 feat: added demos
• 0268588 feat: added full swe bench eval
• bd0b1d5 feat: added logs
• 5c83c8d feat: added max model len
• a624cab feat: added memory manager
• aa6f32f feat: added ollama metrics collection
• 49f0391 feat: added ollama support to demo
• 4dbe34f feat: added one bit stage mapping
• 73c0e54 feat: added orla demo parts
• 6bbfd89 feat: added sglang flushing
• 507b53a feat: added stuck message fix
• 5015992 feat: added support for structured response in open ai api
• 498b428 feat: added tool parser for sglang
• 1ddf54b feat: added tools and updated baseline
• 03d5a1e feat: added two stage mapping experiment
• 42ec3b0 feat: added unimplemented error for structured output in ollama api
• eddaea2 feat: added unit tests for tools
• e57687f feat: adding examples
• a5afe61 feat: better logging for experiments
• 17a75a8 feat: better system prompt
• ec49a58 feat: fixed abstractions end to end
• 6db7da7 feat: fixed patch output error
• 73d610f feat: fixed thinking
• 93eb0bd feat: fixing swe bench eval
• 4271ba4 feat: gpu mem util cap for vllm
• 0570dac feat: improvements inspired by mini_swe_agent
• c37dfb9 feat: logging mapped stage
• 6c7c362 feat: loop detection and gracefull fail
• e3273af feat: made run target configurable
• 0918486 feat: more run target log
• 8ba00ce feat: more swe agent alignment
• ee2dd64 feat: more swebench baseline work
• 05a0991 feat: more swebench work
• 5abca20 feat: name change to my fav mathematician
• 752cf4b feat: new experiments added
• b974f2e feat: parallel backend runs
• 5aff585 feat: reduced mem fraction to support both backends
• 4031eae feat: removed ollama provider
• c1f6359 feat: router prompt engineering
• 319447c feat: running full swebenchlite experiment
• 2c65c7a feat: scheduling and workflow
• a64e141 feat: set temp to 0 for reproducibility
• d208a86 feat: sped up workdir gen for swe-bench
• e169bf1 feat: supporting continuous batching in orla
• 3432b72 feat: suppressed thinking in workflow demo
• 0650770 feat: swebench baseline
• 2071f33 feat: ttft in dag-math eval
• 4d84b7c feat: updated
• c11a46e feat: updated dag math
• bf18635 feat: updated dag math
• a0729f3 feat: updated dag math
• 10c06aa feat: updated demo
• 5445240 feat: updated demo script
• b1015cc feat: updated demo script
• ab24071 feat: updated example
• 194bf2c feat: updated example
• 61bd256 feat: updated light model
• 3c2a805 feat: updated light model
• deeb7f6 feat: updated light weight model back to qwen3 4B
• 0a40f19 feat: updated memory manager
• 533a418 feat: updated memory weights
• df1edb1 feat: updated model
• 2be1083 feat: updated model for dag-math
• 3958ed8 feat: updated one bit router prompt
• 9665946 feat: updated order for dag-math
• 60ec41a feat: updated swe bench eval
• 7676df9 feat: updated swe-bench examples
• 11c5afc feat: updated workflow demo
• 3494313 feat: updated workflow demo
• 118d58a feat: updating demo
• 4cdc8ca feat: updating routing
• 783e98e feat: use light model for stage mapping
• 005212e feat: visible tool calling
• e6aa0ad fix: bash tool context truncation
• ba69e31 fix: bash tool context truncation
• 878dff5 fix: broken unit test
• ef299d4 fix: ci lint
• ddf3ac7 fix: context length issues
• ece1d1a fix: exec target mapping
• b04a896 fix: formatting
• f8dee82 fix: go mod dep
• 163abab fix: lint
• 2be12b8 fix: lint exclusion
• 5c8d742 fix: lint for bash tool
• 925bafb fix: lint for the demo
• 12d3236 fix: linting issues with gosec
• c611d6b fix: mini swe agent
• d0ab358 fix: more nudge handling
• 181120f fix: more workdir creation
• 5151343 fix: more workdir creation
• 9810efe fix: more workdir creation
• 98c4109 fix: more workdir creation
• 52b940c fix: nasty go routine bug
• ec95db6 fix: no sudo in demo
• 17c4754 fix: prompt update
• 64513da fix: removed snippet from test
• 51e0e78 fix: sglang backend is open ai api compatible
• 2e4d141 fix: streaming memory management
• 1fc9e88 fix: swebench lite orla
• 8ee440b fix: tool parser
• 648b50f fix: updated experiment
• ce3f7f3 fix: updated pyproject
• eae718f fix: updated pyproject
• 1d90ef5 fix: weird loop in model loading
• 8a873a9 fix: workdir creating
• a554efb refactor: make snippets templates
• 9d9b879 refactor: modified abstractions
• 4efde4b using alternate light model

---

Full Changelog: v1.2.7...v1.2.8

v1.2.7

Orla v1.2.7

Release v1.2.7 of Orla.

Changelog

• 4aba085 fix: install script

---

Full Changelog: v1.2.6...v1.2.7