8c8dfa482f
feat(server): surface aggregated usage totals to clients
...
Source TokenState and the ACP usage updates from the on-the-fly aggregation
(session + sub-agent tree) rather than the stored session record, so clients
see the derived per-message totals including delegated sub-agent spend.
2026-07-02 09:06:01 -07:00
518e5e7ec8
feat(agents): capture per-message usage in the reply loop
...
Resolve each turn's best-estimate cost (provider-reported when present, else
catalog) once in update_session_metrics and record it to the ledger atomically
with the session totals - for normal turns, auto-compaction, and /compact.
Link sub-agent sessions to their parent so their usage rolls into the parent's
grand total. Attach the turn's usage to its assistant message for display.
2026-07-02 09:06:01 -07:00
d7a7fcc52e
feat(session): per-message usage ledger with on-the-fly aggregation
...
Add an append-only usage_ledger table (migration v15) and a persisted
sessions.parent_session_id link. The ledger is decoupled from the messages
table so per-call usage survives conversation rewrites (compaction) and
cancelled turns, and each row records the model that produced it so estimated
costs can be recomputed or broken down per model later.
Session totals are derived by summing the ledger across a session and its
sub-agent descendants (recursive CTE over parent_session_id), with a
per-session fallback to the retained accumulated_* record for sessions that
predate the migration.
record_usage_metrics runs in one atomic transaction: it first reconciles the
ledger with any spend recorded outside it (pre-v15 history, or turns run on a
pre-v15 build after switching branches) via a 'carried_forward' adjustment row
- both builds only ever grow accumulated_*, so the componentwise difference is
exactly the spend the ledger missed - then adds the call's delta to the
accumulated columns in SQL (so concurrent writers can't lose updates) and
appends the ledger row. The derived total and the legacy fallback therefore
converge again after any branch round-trip.
2026-07-02 09:06:01 -07:00
7028d15e02
feat(conversation): add per-message MessageUsage carrier
...
Add MessageUsage (tokens, cost, cost_source, generation timing) on
MessageMetadata, populated and serialized to clients as groundwork for the
per-message UI, and register it plus CostSource in the OpenAPI schema.
2026-07-02 09:06:01 -07:00
71095b16c7
feat(providers): record provider-reported cost on ProviderUsage
...
Add optional cost + cost_source (ProviderReported/Estimated) to ProviderUsage
and parse `usage.cost` from OpenRouter and Anthropic-compatible gateways across
the streaming and non-streaming OpenAI and Anthropic paths. Request OpenRouter's
inline cost via `usage: {include: true}`. When no provider cost is present the
cost is filled in downstream from the catalog estimate.
Drop the unused ProviderUsage::combine_with rather than extending it with cost
semantics nothing exercises.
2026-07-02 09:06:01 -07:00
79cbb2627d
Fail closed on unverifiable update provenance ( #10168 )
...
Co-authored-by: Jasper <jasper@spiral.xyz >
2026-07-02 15:09:38 +00:00
402ff562b3
chore(release): bump version to 1.41.0 (minor) ( #10198 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Lifei Zhou <lifei@squareup.com >
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com >
2026-07-02 12:41:48 +00:00
a3854aee3d
Fix canonical model output paths ( #10196 )
2026-07-02 11:05:16 +00:00
ba6fb2b56d
Refactor local inference provider crates ( #10169 )
2026-07-02 08:51:53 +00:00
bf1a986d5e
Allow local file ACP origins ( #10194 )
2026-07-02 08:37:06 +00:00
410a78983e
fix(acp): propagate agent cleanup errors on session delete ( #10112 )
...
Signed-off-by: Abhijay Jain <Abhijay007j@gmail.com >
2026-07-02 08:06:41 +00:00
dbc1f88ca7
fix(providers): add claude-sonnet-5 to canonical model registry ( #10183 )
...
Co-authored-by: syf2211 <syf2211@users.noreply.github.com >
2026-07-02 06:10:35 +00:00
185b185fb8
fix: increased mcp-app-guest request body limit ( #10180 )
2026-07-02 04:30:53 +00:00
b83b194dad
acp migration (ui) : ui connect to acp directly instead of goosed ( #10081 )
2026-07-02 00:29:29 +00:00
0d4b6924b9
Revert "Bedrock model discovery and validation" ( #10176 )
2026-07-01 23:39:04 +00:00
a162a7f778
Improve ACP auth and origin defaults ( #9886 )
2026-07-01 22:08:46 +00:00
7026e14585
Bedrock model discovery and validation ( #9997 )
2026-07-01 21:19:28 +00:00
9165e2738f
fix: tolerate partial Responses output items ( #10108 )
2026-07-01 19:56:57 +00:00
44869db2d9
fix(databricks): always use OpenAI-compatible "parameters" for tools ( #10101 )
2026-07-01 19:56:54 +00:00
2d47d1feea
fix(providers): unescape shell-escaped image paths ( #10098 )
2026-07-01 19:56:52 +00:00
a3b4166d4e
Update EmpirioLabs provider display name to EmpirioLabs AI ( #10126 )
2026-07-01 17:06:38 +00:00
b6349b5125
add declarative provider support to goose-providers crate ( #9992 )
2026-07-01 16:40:49 +00:00
f150672e23
test: generic validator for declarative providers ( #10010 )
...
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
2026-07-01 03:13:10 +00:00
66ce9a6beb
Remove MCP sampling support ( #10087 )
...
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
Co-authored-by: Lifei Zhou <lifei@squareup.com >
2026-07-01 01:53:44 +00:00
c8971fe5fb
Support TLS for ACP serve ( #10088 )
...
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
Co-authored-by: Lifei Zhou <lifei@squareup.com >
2026-07-01 00:43:18 +00:00
2f32070975
feat (ui): Migrate dictation local model manager to ACP ( #10131 )
2026-07-01 00:02:52 +00:00
e8891b7bb0
fix(bedrock): send inference config (max_tokens, temperature) on Converse ( #9889 )
...
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com >
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
2026-06-30 22:59:40 +00:00
12c3e234f5
feat(providers): support OpenRouter request parameters ( #9276 )
...
Signed-off-by: sunilkumarvalmiki <g.sunilkumarvalmiki@gmail.com >
Signed-off-by: Douwe Osinga <douwe@sidewalklabs.com >
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
2026-06-30 22:48:03 +00:00
d1cc1439ec
Migrate local inference model management to ACP ( #10124 )
2026-06-30 22:40:04 +00:00
20966fd057
feat: add --edit session flag to edit conversation before forking ( #9799 )
...
Signed-off-by: Sean Murphy <murphysean84@gmail.com >
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
2026-06-30 21:15:52 +00:00
04da8e8702
feat: add iFlytek Spark and Astron MaaS providers ( #9837 )
...
Signed-off-by: FenjuFu <fufenjupku@gmail.com >
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
2026-06-30 21:15:50 +00:00
d308b13e75
[codex] Add SessionStart hook parity outside CLI ( #9970 )
...
Co-authored-by: Max Novich <mnovich@Maxs-MacBook-Pro.local >
2026-06-30 21:15:46 +00:00
40bfb137ad
feat(providers): add Fireworks AI declarative provider ( #9990 )
...
Signed-off-by: Leo Walker <leowalker89@gmail.com >
2026-06-30 21:15:43 +00:00
c45f757c9d
fix(providers): don't retry deterministically-permanent 400s (thinking-block immutability) ( #10005 )
...
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
2026-06-30 21:15:41 +00:00
d056c72a2c
fix(deps): downgrade pkcs8 to v0.10 to match sec1/pkcs1 v0.7 ( #10119 )
...
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com >
2026-06-30 21:15:38 +00:00
a90430e8e5
Make OpenAI Responses API store param configurable ( #10040 )
...
Signed-off-by: Lucas Alvares Gomes <lucasagomes@gmail.com >
2026-06-30 20:00:12 +00:00
f1a97fde83
chore(release): bump version to 1.40.0 (minor) ( #10099 )
...
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Jack Amadeo <jackamadeo@squareup.com >
2026-06-30 18:15:28 +00:00
b806204887
move ollama provider into goose-providers ( #9986 )
2026-06-30 17:28:45 +00:00
2084c43476
feat (acp): exposed available tools in acp schema ( #10097 )
2026-06-30 05:43:08 +00:00
d43c692127
perf(providers): keep turn-context out of the Anthropic prefix cache ( #10030 )
2026-06-30 05:09:37 +00:00
37e62497d4
feat (acp+): Remove REST onboarding setup paths ( #10061 )
2026-06-30 00:35:30 +00:00
4778448b49
fix(test): stop flaky unparseable-tool-call test by disabling session naming ( #10091 )
2026-06-30 00:12:07 +00:00
f81a0e2471
chore: Remove legacy MCP-UI proxy support ( #10086 )
2026-06-29 22:58:40 +00:00
fe7f16b727
Remove session_id from provider streaming trait methods ( #9984 )
2026-06-29 22:04:58 +00:00
84d02d6d47
fix(cli): update help text for --session-id ( #10077 )
...
Signed-off-by: Vikram Vaswani <2571660+vvaswani@users.noreply.github.com >
2026-06-29 15:30:27 +00:00
2cc1140dc1
task(acp): upgrade SDK and use new HTTP/WS crate ( #10082 )
2026-06-29 05:27:11 +00:00
a1711e75e2
fix(providers): reject non-object tool-call arguments instead of panicking ( #9832 )
...
Signed-off-by: Lunchbox <lunchboxfortwo@gmail.com >
Signed-off-by: Douwe M Osinga <douwe@sidewalklabs.com >
Signed-off-by: Goose <goose@aaif.dev >
Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com >
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
Co-authored-by: Goose <goose@aaif.dev >
Co-authored-by: Michael Neale <michael.neale@gmail.com >
2026-06-29 04:08:19 +00:00
71280996d4
fix(cli): save /edit prompts to history ( #10011 )
...
Co-authored-by: Douwe M Osinga <douwe@sidewalklabs.com >
2026-06-28 16:44:01 +00:00
52291e64a7
feat (acp+): Use ACP permission manager for tool permissions ( #10066 )
2026-06-28 09:54:51 +00:00
d2687643da
fix(schedule): use session.message_count for schedule sessions listing ( #10026 )
...
Co-authored-by: Cursor Agent <cursoragent@cursor.com >
2026-06-27 16:24:54 +00:00
Filip Kujawa