Files
ocp/ocp-connect
T
3dc023e596 feat(models): align maxTokens with the CLI registry (#195) (#208)
* feat(models): align maxTokens with the CLI registry (#195)

Every Opus entry and claude-sonnet-5 go 16384 -> 64000; claude-sonnet-4-6 and
claude-haiku-4-5 go to 32000 (haiku was 8192). These are the
max_output_tokens.default each model declares in the compiled CLI 2.1.220
registry, extracted id-anchored:

  claude-opus-5      max_output_tokens:{default:64000,upper:128000}
  claude-opus-4-8    max_output_tokens:{default:64000,upper:128000}
  claude-opus-4-7    max_output_tokens:{default:64000,upper:128000}
  claude-opus-4-6    max_output_tokens:{default:64000,upper:128000}
  claude-sonnet-5    max_output_tokens:{default:64000,upper:128000}
  claude-sonnet-4-6  max_output_tokens:{default:32000,upper:128000}
  claude-haiku-4-5   max_output_tokens:{default:32000,upper:64000}

This is not cosmetic metadata. OCP itself never enforces maxTokens — server.mjs
does not read it — but it is propagated to OpenClaw (setup.mjs,
scripts/sync-openclaw.mjs), and OpenClaw CAPS the outbound request with it:

  maxTokens = Math.min(baseMaxTokens + thinkingBudget, modelMaxTokens)

with reasoning budgets of medium 8192 / high 16384, and a clamp
`if (maxTokens <= thinkingBudget) thinkingBudget = maxTokens - 1024`.

So at `high` reasoning against a model declared at 16384: maxTokens resolves to
16384, the clamp fires, thinking takes 15360, and roughly 1024 tokens are left
for the visible answer. That is what this repo shipped for every model until
now, and it is the actual harm — not the advertised-vs-real mismatch.

ocp-connect keeps its own family-prefix table (/v1/models does not expose
maxTokens, so it cannot derive them). Set to the family FLOOR — opus 64000,
sonnet 32000 because sonnet-4-6 is 32000 while sonnet-5 is 64000, haiku 32000 —
because prefixes cannot distinguish versions and under-advertising merely caps
a client lower than the model allows, whereas over-advertising would promise
capacity a family member does not have. Commented in place.

New test asserts the PRINCIPLE, not the numbers: every maxTokens must exceed
OpenClaw's `high` thinking budget so an answer still fits. A value-pinning test
would need editing on every model addition and would not explain itself.
Mutation-proven: reverting one entry to 16384 fails with
"claude-opus-5 declares maxTokens=16384 <= 16384: at OpenClaw's 'high'
reasoning level the thinking budget would consume all but ~1024 tokens".

Owner-approved decision (option: align with registry defaults). Tradeoff
recorded in CHANGELOG: longer answers, correspondingly higher per-request quota
burn on long generations.

Verified: bash -n on ocp-connect, and py_compile on its embedded python block
(lines 101-334) so the comment sits at a valid indentation.

Tests: 458 passed, 0 failed (457 + 1).

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017gbqUZ8HfBZpjjbzQ85oH8

* fix(models): rewrite the maxTokens rationale — my justification was factually wrong

The numbers were right, the code was right, and the reason I gave for both was
wrong. Verified every part of the review's finding first-hand before rewriting.

WHAT I CLAIMED: OpenClaw caps outbound requests via
`maxTokens = Math.min(baseMaxTokens + thinkingBudget, modelMaxTokens)`, so a
16384 declaration at `high` reasoning left ~1024 tokens for the visible answer,
and raising it would produce longer answers and higher quota burn.

WHY THAT CANNOT HAPPEN FOR OCP — three independent breaks, each confirmed here:

  scripts/sync-openclaw.mjs:75   api: "openai-completions"
The clamp I cited lives behind `case "anthropic-messages"`; the
openai-completions transport has no thinkingBudget concept at all, so that code
never runs for a claude-local provider.

  buildCliArgs pushes: --input-format, --disallowedTools, --allowedTools,
                       --dangerously-skip-permissions, --mcp-config
No output-token flag of any kind reaches the CLI. OCP cannot cap output even if
it wanted to.

  grep max_completion_tokens server.mjs lib/  ->  0 occurrences
That is the field OpenClaw would send to a local openai-completions endpoint,
and this repo never reads it.

So "expect longer answers and higher per-request quota burn" was a promise that
would not have materialised. That is worse than a wrong number: a release note
that tells operators to expect a behavioural change which cannot occur.

The change still stands, on the honest reason: models.json is the SPOT (ADR
0003) and its values should be true. maxTokens is ADVERTISED metadata consumed
only by clients that choose to honour it. CHANGELOG and the test comment now
say exactly that, and say plainly that OCP behavior is unchanged.

Test replaced, not just re-worded. The old one asserted `maxTokens > 16384`,
which review showed lets every entry sit at 16385 and still call itself
"registry-aligned" — the actual claim had no coverage. Now pinned per model
against a recorded table of CLI 2.1.220 id-anchored values, with an explicit
failure for any model missing a row. Mutation-proven on both holes:

  all entries at 16385          -> CAUGHT  (the old test passed this)
  sonnet-4-6 wrongly at 64000   -> CAUGHT

Also fixed ocp-connect's unknown-id fallback (8192 -> 32000). Review flagged it
as contradicting the invariant: 32000 is the LOWEST max_output_tokens.default
in the registry, so it is the safe floor and no longer sits below every family
in its own table. Unreachable today, wrong to leave inconsistent.

Verified: bash -n on ocp-connect and py_compile on its embedded python block.
Tests: 458 passed, 0 failed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017gbqUZ8HfBZpjjbzQ85oH8

* fix(ocp-connect): revert the unknown-id fallback to 8192 — 32000 over-advertised

The previous commit raised ocp-connect's unknown-id fallback from 8192 to 32000 on
the stated grounds that 32000 was "the LOWEST max_output_tokens.default in the CLI
2.1.220 registry". That is false, and the change it justified made this PR violate
its own thesis.

Pairing-free enumeration of the compiled registry (17 occurrences, no grep-window
artifacts):

    7 max_output_tokens:{default:64000,upper:128000}
    5 max_output_tokens:{default:32000,upper:64000}
    2 max_output_tokens:{default:8192,upper:8192}
    2 max_output_tokens:{default:32000,upper:32000}
    1 max_output_tokens:{default:32000,upper:128000}

    distinct defaults: 8192 / 32000 / 64000     GLOBAL MINIMUM: 8192

The floor is 8192, held by claude-3-5-haiku and claude-3-5-sonnet — and those are
exactly the ids that reach the fallback, since "claude-3-5-*" matches none of the
family prefixes. So the raise over-advertised the only two real models the branch
serves, by 4x, in a PR whose entire claim is that advertised metadata should be true.

Verified by exercising the real classifier (verbatim source slice, not a replica):

    claude-3-5-haiku            8192  FALLBACK   registry 8192   exact
    claude-3-5-sonnet           8192  FALLBACK   registry 8192   exact
    claude-opus-5              64000  family     registry 64000  exact
    claude-sonnet-4-6          32000  family     registry 32000  exact
    claude-haiku-4-5-20251001  32000  family     registry 32000  exact
    claude-sonnet-5            32000  family     registry 64000  under (safe)

Also in this commit:

- The family table's comment said "FAMILY FLOOR of the CLI registry". It is the floor
  over each family's members CURRENTLY IN models.json, not over the registry family:
  claude-opus-4-0 / -4-5 are opus-family at 32000, so the opus row would over-advertise
  them 2x. Unreachable today (ocp-connect derives ids from /v1/models, i.e. models.json),
  but the comment now states the real scope and what to do if a legacy opus is added.

- The test table keyed haiku on "claude-haiku-4-5-20251001" while telling the reader to
  extract values id-anchored. That id has ZERO id-anchored hits: the registry record is
  id:"claude-haiku-4-5" and the dated string appears only as its provider_ids.first_party
  (10 bare-string hits). Following the instruction on that row returns nothing, which is
  precisely what pushes a reader back to the bare-string search this test exists to
  prevent. The row now carries the registry id.

Reviewer credit: the reviewer that found this also retracted its own earlier finding,
which is what produced the 32000 fallback in the first place. Both directions verified
independently here before acting.

server.mjs: unchanged by this commit (ALIGNMENT.md requires no cli.js citation).
Suite: 462 passed, 0 failed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017gbqUZ8HfBZpjjbzQ85oH8

* docs(ocp-connect): my unreachability reason was wrong — state the invariant that holds

Review of cb87b82 approved the change and flagged the new comment's reasoning. Verified
both entry points it named; both are real.

I wrote that claude-opus-4-0/-4-5 are unreachable because "ocp-connect derives ids from
/v1/models, i.e. models.json". Neither half is reliable:

- base_url is user-supplied (ocp-connect:510 builds it from arguments), so /v1/models is
  the REMOTE server's models.json — a different OCP version, or a fork, not this checkout.
- on a JSON parse failure the ids come from a hardcoded three-id list (:114), not from
  /v1/models at all. Checked all three: opus-4-6 -> 64000, sonnet-4-6 -> 32000,
  haiku-4 -> 32000. All exact against the registry, so safe today — but it is a genuine
  non-SPOT entry point that my sentence denied existed.

The invariant that actually holds is narrower: no released OCP has ever served a legacy
opus id. Same conclusion, and now it names the thing that would break it.

This is the second reasoning defect in this PR where the number was right and the
justification was not. Recording it in the comment rather than quietly fixing it.

Out of scope, noted for a future touch: the bare `except:` at :114 catches BaseException,
so a KeyboardInterrupt mid-parse silently yields the fallback list.

server.mjs: unchanged (ALIGNMENT.md requires no cli.js citation).
Suite: 462 passed, 0 failed.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017gbqUZ8HfBZpjjbzQ85oH8

---------

Co-authored-by: dtzp555 <dtzp555@gmail.com>
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
2026-07-27 12:26:51 +10:00

759 lines
29 KiB
Bash
Executable File

#!/usr/bin/env bash
# ocp-connect — Lightweight client script to connect to a remote OCP instance
# No dependencies beyond curl and python3 (available on most Linux/Mac systems)
#
# Install:
# curl -fsSL https://raw.githubusercontent.com/dtzp555-max/ocp/main/ocp-connect -o ocp-connect
# chmod +x ocp-connect
#
# Or run directly:
# curl -fsSL https://raw.githubusercontent.com/dtzp555-max/ocp/main/ocp-connect | bash -s -- <host-ip> --key <key>
#
set -euo pipefail
OCP_CONNECT_VERSION="1.3.0"
show_version() {
echo "ocp-connect $OCP_CONNECT_VERSION"
}
show_help() {
cat <<'EOF'
ocp-connect — Connect this machine to a remote OCP (Open Claude Proxy)
Configures OPENAI_BASE_URL and OPENAI_API_KEY in your shell rc file
so tools like Claude Code, Cline, Aider, etc. point to the remote OCP.
Usage:
ocp-connect <host-ip> [options]
Options:
--port PORT Port OCP listens on (default: 3456)
--key API_KEY API key (prompted interactively if remote requires auth)
--version Print version and exit
--help, -h Show this help
Examples:
ocp-connect 192.168.1.10
ocp-connect 192.168.1.10 --port 8080
ocp-connect 192.168.1.10 --key ocp_abc123
What it does:
1. Tests connectivity to the remote OCP
2. Verifies your API key (if auth is enabled)
3. Writes OPENAI_BASE_URL and OPENAI_API_KEY to ~/.bashrc or ~/.zshrc
4. Sets system-level env vars (launchctl on macOS, systemd on Linux)
5. Configures OpenClaw automatically; prints setup hints for other IDEs
(Cline, Continue.dev, Cursor — manual configuration required)
6. Runs a smoke test to confirm everything works
After running, reload your shell: source ~/.bashrc (or ~/.zshrc)
EOF
}
configure_ides() {
local base_url="$1" key="$2" models_out="$3"
# --- OpenClaw ---
local oc_config="$HOME/.openclaw/openclaw.json"
local oc_found=false
if command -v openclaw &>/dev/null || [[ -f "$oc_config" ]]; then
oc_found=true
fi
if $oc_found; then
echo " IDE Configuration"
echo " ─────────────────────────────────────"
echo " Detected: OpenClaw ($oc_config)"
echo ""
printf " Configure OpenClaw to use this OCP? [Y/n] "
local oc_answer
{ read -r oc_answer </dev/tty; } 2>/dev/null || oc_answer="y"
oc_answer="${oc_answer:-y}"
if [[ "$oc_answer" =~ ^[Yy]$ ]]; then
# Ask for provider name
printf " Provider name (models show as <name>/model-id) [ocp]: "
local provider_name
{ read -r provider_name </dev/tty; } 2>/dev/null || provider_name=""
provider_name="${provider_name:-ocp}"
# Sanitize: only allow alphanumeric, dash, underscore
provider_name=$(echo "$provider_name" | tr -cd 'a-zA-Z0-9_-')
[[ -z "$provider_name" ]] && provider_name="ocp"
# Ask for priority
echo ""
echo " How should OCP models be configured?"
echo " 1) Primary — use OCP by default, keep existing models as backup"
echo " 2) Backup — keep current primary, add OCP as additional option"
echo ""
printf " Choice [1]: "
local priority_choice
{ read -r priority_choice </dev/tty; } 2>/dev/null || priority_choice="1"
priority_choice="${priority_choice:-1}"
echo ""
echo " Writing OpenClaw config..."
# Use python3 to safely manipulate JSON
local py_ok=0
python3 - "$oc_config" "$base_url" "$key" "$provider_name" "$priority_choice" "$models_out" <<'PYEOF' && py_ok=1
import sys, json, os
config_path = sys.argv[1]
base_url = sys.argv[2]
api_key = sys.argv[3]
provider_name = sys.argv[4]
priority = sys.argv[5] # "1" = primary, "2" = backup
models_json_str = sys.argv[6]
# Parse models from OCP /v1/models response
try:
models_data = json.loads(models_json_str)
model_ids = [m["id"] for m in models_data.get("data", [])]
except:
model_ids = ["claude-opus-4-6", "claude-sonnet-4-6", "claude-haiku-4"]
# Build provider entry
provider = {
"baseUrl": base_url + "/v1",
"api": "openai-completions",
"authHeader": bool(api_key),
"models": []
}
# Model metadata mapping. Prefix match on the model FAMILY (claude-opus / -sonnet /
# -haiku), not a pinned version. A version-pinned prefix like "claude-sonnet-4"
# silently misses "claude-sonnet-5" and falls through to the non-reasoning /
# 8k-output default (PR #152 review) — every future Sonnet/Opus/Haiku bump would
# re-trip it. Family prefixes classify any versioned ID correctly with no per-model
# edit. (ADR 0003: models.json is the SPOT for model existence; /v1/models does not
# expose reasoning/maxTokens, so family classification stays here.)
# maxTokens values are the floor of the CLI registry's max_output_tokens.default over the
# family members CURRENTLY IN models.json (opus 64000; sonnet 32000, because sonnet-4-6 is
# 32000 while sonnet-5 is 64000; haiku 32000) — NOT over the whole registry family. The
# registry also holds claude-opus-4-0 / -4-5 at 32000, which this opus row would over-advertise
# 2x. What keeps that unreachable is NOT "the ids come from models.json" — the ids reaching
# get_model_meta have two sources that are not this checkout: /v1/models is fetched from a
# user-supplied base_url, so it is the REMOTE server's models.json (possibly a different OCP
# version or a fork), and on a JSON parse failure the ids come from the hardcoded list a few
# lines above instead. The invariant that actually holds is that no released OCP has ever
# served a legacy opus id. Adding one to models.json breaks it and this row must drop to 32000.
# Family prefixes cannot distinguish versions, so a floor is used deliberately:
# under-advertising caps a client lower than the model allows, which is safe, whereas
# over-advertising would promise capacity a member of the family does not have.
# models.json is authoritative per ADR 0003; this table only exists because /v1/models does
# not expose maxTokens.
model_meta = {
"claude-opus": {"name": "Claude Opus (OCP)", "reasoning": True, "maxTokens": 64000},
"claude-sonnet": {"name": "Claude Sonnet (OCP)", "reasoning": True, "maxTokens": 32000},
"claude-haiku": {"name": "Claude Haiku (OCP)", "reasoning": False, "maxTokens": 32000},
}
def get_model_meta(mid):
"""Match model metadata by prefix."""
for prefix, meta in sorted(model_meta.items(), key=lambda x: -len(x[0])):
if mid.startswith(prefix):
return meta
# Unknown id. 8192 is the GLOBAL MINIMUM max_output_tokens.default across all 17 records in
# the CLI 2.1.220 registry (distinct values: 8192 / 32000 / 64000). The two records holding
# it — claude-3-5-haiku and claude-3-5-sonnet, both default:8192, upper:8192 — are precisely
# the ones that reach this line, because "claude-3-5-*" matches none of the family prefixes
# above. Raising this to a family-table value would over-advertise those two by 4x. Keep 8192:
# under-advertising caps a client lower than a model allows, which is safe; over-advertising
# promises capacity that does not exist.
return {"name": mid + " (OCP)", "reasoning": False, "maxTokens": 8192}
for mid in model_ids:
meta = get_model_meta(mid)
provider["models"].append({
"id": mid,
"name": meta["name"],
"reasoning": meta["reasoning"],
"input": ["text"],
"cost": {"input": 0, "output": 0, "cacheRead": 0, "cacheWrite": 0},
"contextWindow": 200000,
"maxTokens": meta["maxTokens"]
})
# Load or create config
if os.path.exists(config_path):
with open(config_path, "r") as f:
config = json.load(f)
else:
os.makedirs(os.path.dirname(config_path), exist_ok=True)
config = {}
# Ensure models.providers exists
config.setdefault("models", {})
config["models"].setdefault("mode", "merge")
config["models"].setdefault("providers", {})
# Remove any previous OCP provider with the same name
config["models"]["providers"][provider_name] = provider
# Set up auth profile if key is provided
if api_key:
config.setdefault("auth", {})
config["auth"].setdefault("profiles", {})
config["auth"]["profiles"][provider_name + ":default"] = {
"provider": provider_name,
"mode": "api_key"
}
# Configure agent defaults — add model aliases
config.setdefault("agents", {})
config["agents"].setdefault("defaults", {})
config["agents"]["defaults"].setdefault("models", {})
# Build alias map (family prefix match — version-agnostic, see model_meta note)
alias_prefixes = {
"claude-opus": "Claude Opus",
"claude-sonnet": "Claude Sonnet",
"claude-haiku": "Claude Haiku",
}
for mid in model_ids:
full_id = provider_name + "/" + mid
alias = mid # fallback
for prefix, name in sorted(alias_prefixes.items(), key=lambda x: -len(x[0])):
if mid.startswith(prefix):
alias = name
break
config["agents"]["defaults"]["models"][full_id] = {"alias": alias}
# Handle primary/backup
if priority == "1":
# OCP as primary — pick the best model (prefer the latest Sonnet for daily use,
# tracking the `sonnet` alias default in models.json; fall back across versions).
_sonnet_pref = ["claude-sonnet-5", "claude-sonnet-4-6"]
primary_model = next(
(provider_name + "/" + m for m in _sonnet_pref if m in model_ids),
provider_name + "/" + model_ids[0],
)
config["agents"]["defaults"].setdefault("model", {})
config["agents"]["defaults"]["model"]["primary"] = primary_model
# Keep existing fallbacks
config["agents"]["defaults"]["model"].setdefault("fallbacks", [])
# If backup (priority == "2"), don't change the primary — just add models to the list
# Update agent list entries that use old provider name patterns
# (only update if agents.list exists and has entries using old OCP-like providers)
if "list" in config.get("agents", {}):
for agent in config["agents"]["list"]:
agent_model = agent.get("model", {})
if priority == "1" and agent_model.get("primary", "").startswith("claude-local/"):
# Migrate from claude-local to new provider
old_model_id = agent_model["primary"].split("/", 1)[1]
if old_model_id in model_ids:
agent_model["primary"] = provider_name + "/" + old_model_id
# Also update subagents if they use claude-local
sub = agent.get("subagents", config["agents"]["defaults"].get("subagents", {}))
# Don't modify subagents in agent entries — they inherit from defaults
# Update defaults subagents model if using claude-local
if priority == "1":
sub = config["agents"]["defaults"].get("subagents", {})
if sub.get("model", "").startswith("claude-local/"):
old_id = sub["model"].split("/", 1)[1]
if old_id in model_ids:
sub["model"] = provider_name + "/" + old_id
with open(config_path, "w") as f:
json.dump(config, f, indent=2, ensure_ascii=False)
f.write("\n")
# === B1 fix: seed per-agent auth-profiles.json for OpenClaw multi-agent setups ===
# OpenClaw's per-agent auth loader reads <agentDir>/auth-profiles.json (NOT the
# root openclaw.json's auth.profiles section). Without a real key in each agent's
# agentDir, OpenClaw rejects the lane with "No API key found for provider X".
# See https://github.com/dtzp555-max/ocp/issues/12 for the full investigation.
_seeded = []
_failed = []
_skipped_anonymous = []
_profile_key = provider_name + ":default"
# OpenClaw stores per-agent auth in <openclaw_root>/agents/<id>/agent/ when
# the agent has no explicit `agentDir` field — derive that path so the
# default `main` agent (which never sets agentDir) is also seeded.
_openclaw_root = os.path.dirname(config_path)
for _agent in config.get("agents", {}).get("list", []):
_agent_dir = _agent.get("agentDir")
if not _agent_dir:
_agent_id = _agent.get("id")
if not _agent_id:
continue
_agent_dir = os.path.join(_openclaw_root, "agents", _agent_id, "agent")
if not api_key:
# Anonymous mode is incompatible with OpenClaw per-agent auth (empty key
# is dropped by OpenClaw's pi-auth-credentials). Per OCP issue #12 §14
# decision: take Path C (require --key for OpenClaw multi-agent setups).
_skipped_anonymous.append(_agent_dir)
continue
_profiles_path = os.path.join(_agent_dir, "auth-profiles.json")
try:
os.makedirs(_agent_dir, exist_ok=True)
except OSError as _e:
_failed.append((_profiles_path, "mkdir: " + str(_e)))
continue
if os.path.exists(_profiles_path):
try:
with open(_profiles_path) as _f:
_ap = json.load(_f)
except json.JSONDecodeError:
# Corrupted JSON — back up and rebuild from scratch.
_bak = _profiles_path + ".bak"
try:
os.rename(_profiles_path, _bak)
print(f" ⚠ {_profiles_path} was corrupt; backed up to {_bak}")
except OSError:
pass
_ap = {"version": 1, "profiles": {}}
except OSError as _e:
# Read failure (permissions, disk) — skip to AVOID clobbering
# the user's existing profile (which may hold other providers' keys).
_failed.append((_profiles_path, "read: " + str(_e)))
continue
else:
_ap = {"version": 1, "profiles": {}}
_ap.setdefault("version", 1)
_ap.setdefault("profiles", {})
_ap["profiles"][_profile_key] = {
"type": "api_key",
"provider": provider_name,
"key": api_key
}
try:
with open(_profiles_path, "w") as _f:
json.dump(_ap, _f, indent=2, ensure_ascii=False)
_f.write("\n")
os.chmod(_profiles_path, 0o600)
_seeded.append(_profiles_path)
except OSError as _e:
_failed.append((_profiles_path, "write: " + str(_e)))
# Report — all three sections are independent (mixed scenarios are surfaced).
if _seeded:
print(f" ✓ Per-agent auth profile seeded ({len(_seeded)}):")
for _p in _seeded:
print(f" • {_p}")
if _failed:
print(f" ⚠ Per-agent auth profile FAILED ({len(_failed)}):")
for _p, _err in _failed:
print(f" • {_p}: {_err}")
print(" OpenClaw will report \"No API key found\" for the failed agents.")
print(" Fix the underlying error (permissions / disk) and re-run ocp-connect.")
if _skipped_anonymous:
print(f" ⚠ OpenClaw multi-agent mode detected ({len(_skipped_anonymous)} agents).")
print(" Anonymous mode does not work for OpenClaw per-agent auth.")
print(" Re-run with: ocp-connect <host> --key ocp_xxx")
PYEOF
if [[ $py_ok -eq 1 ]]; then
echo " ✓ OpenClaw configured"
echo " Provider: $provider_name"
echo " Models:"
# List models from the already-fetched models_out
echo "$models_out" | python3 -c "
import sys,json
d=json.loads(sys.stdin.read())
pn='$provider_name'
for m in d.get('data',[]):
print(' • ' + pn + '/' + m['id'])
" 2>/dev/null
if [[ "$priority_choice" == "1" ]]; then
echo " Priority: PRIMARY (default model)"
else
echo " Priority: BACKUP (available in model selector)"
fi
echo ""
echo " Restart OpenClaw to apply: openclaw gateway restart"
else
echo " ⚠ Failed to write OpenClaw config. You can configure it manually."
fi
else
echo " Skipped OpenClaw configuration."
fi
echo ""
fi
# --- Other IDEs: print manual instructions ---
local other_ides_shown=false
# Collect VS Code extension list once (reused by Cline and Continue.dev checks)
local _vscode_exts=""
if command -v code &>/dev/null; then
_vscode_exts=$(code --list-extensions 2>/dev/null || true)
fi
if [[ -z "$_vscode_exts" && -d "$HOME/.vscode/extensions" ]]; then
_vscode_exts=$(ls "$HOME/.vscode/extensions/" 2>/dev/null || true)
fi
# Extract model IDs from /v1/models response for use in hints
local _model_ids
_model_ids=$(echo "$models_out" | python3 -c "
import sys,json
try:
d=json.loads(sys.stdin.read())
print(', '.join(m['id'] for m in d.get('data', [])))
except Exception:
print('claude-sonnet-4-6, claude-opus-4-6, claude-haiku-4-5-20251001')
" 2>/dev/null || echo "claude-sonnet-4-6, claude-opus-4-6, claude-haiku-4-5-20251001")
# Key display: truncate if longer than 16 chars to avoid screenshot leakage,
# show explicit "(none)" in anonymous mode so users don't paste blank fields.
local _key_display
if [[ -z "$key" ]]; then
_key_display="(none — anonymous mode; most external IDEs require a non-empty API Key)"
elif [[ ${#key} -gt 16 ]]; then
_key_display="${key:0:8}...${key: -4}"
else
_key_display="$key"
fi
# Detect Cline (VS Code extension saoudrizwan.claude-dev, see issue #12)
if echo "$_vscode_exts" | grep -qiE 'cline|saoudrizwan\.claude-dev'; then
if ! $other_ides_shown; then
echo " Other IDEs detected:"
other_ides_shown=true
fi
echo " • Cline: VSCode → Cline panel → Settings → API Provider = \"OpenAI Compatible\""
echo " Base URL: $base_url/v1"
echo " API Key: $_key_display"
echo " Model ID: $_model_ids"
fi
# Detect Continue.dev (extension ID continue.continue or config file, see issue #12)
if echo "$_vscode_exts" | grep -qi 'continue\.continue' \
|| [[ -f "$HOME/.continue/config.yaml" ]] \
|| [[ -f "$HOME/.continue/config.json" ]]; then
if ! $other_ides_shown; then
echo " Other IDEs detected:"
other_ides_shown=true
fi
echo " • Continue.dev: edit ~/.continue/config.yaml — add under \`models:\` (top-level key):"
echo " models:"
echo " - name: OCP Sonnet"
echo " provider: openai"
echo " model: claude-sonnet-4-6"
echo " apiBase: $base_url/v1"
echo " apiKey: $_key_display"
echo " (other model IDs: $_model_ids)"
fi
# Detect Cursor (command, ~/.cursor dir, or /Applications/Cursor.app, see issue #12)
if command -v cursor &>/dev/null \
|| [[ -d "$HOME/.cursor" ]] \
|| [[ -d "/Applications/Cursor.app" ]]; then
if ! $other_ides_shown; then
echo " Other IDEs detected:"
other_ides_shown=true
fi
echo " • Cursor: Cmd+Shift+P → 'Cursor Settings' → Models →"
echo " OpenAI API Key: $_key_display"
echo " Override OpenAI Base URL: $base_url/v1"
echo " Custom OpenAI Models: $_model_ids"
fi
# Detect opencode (https://opencode.ai, SST team CLI, see issue #12)
if command -v opencode &>/dev/null \
|| [[ -x "$HOME/.opencode/bin/opencode" ]] \
|| [[ -d "$HOME/.local/share/opencode" ]]; then
if ! $other_ides_shown; then
echo " Other IDEs detected:"
other_ides_shown=true
fi
echo " • opencode: not yet auto-configured by ocp-connect (PR follow-up)."
echo " Run \`opencode providers login openai\` and provide:"
echo " Base URL: $base_url/v1"
echo " API Key: $_key_display"
echo " Available model IDs: $_model_ids"
fi
if $other_ides_shown; then
echo ""
fi
}
main() {
local host="" port=3456 key=""
# Parse args
while [[ $# -gt 0 ]]; do
case "$1" in
--port) port="${2:?'--port requires a value'}"; shift 2 ;;
--key) key="${2:?'--key requires a value'}"
[[ -z "$key" ]] && { echo "Error: --key cannot be empty (omit --key entirely for anonymous mode)"; exit 1; }
shift 2 ;;
--version) show_version; exit 0 ;;
--help|-h) show_help; exit 0 ;;
-*) echo "Unknown option: $1"; show_help; exit 1 ;;
*) host="$1"; shift ;;
esac
done
if [[ -z "$host" ]]; then
echo "Error: host IP is required."
echo ""
show_help
exit 1
fi
if ! [[ "$host" =~ ^[a-zA-Z0-9._-]+$ ]]; then
echo "Error: invalid host '$host'"
exit 1
fi
# Check dependencies
for cmd in curl python3; do
if ! command -v "$cmd" &>/dev/null; then
echo "Error: '$cmd' is required but not found."
exit 1
fi
done
local base_url="http://$host:$port"
echo "OCP Connect v$OCP_CONNECT_VERSION"
echo "─────────────────────────────────────"
echo " Remote: $base_url"
echo ""
# Step 1: Test connectivity via /health
echo " Checking connectivity..."
local health_json
health_json=$(curl -sf --max-time 10 "$base_url/health" 2>/dev/null) || {
echo " ✗ Cannot reach $base_url/health"
echo " Make sure OCP is running on $host and bound to 0.0.0.0 (LAN mode)."
exit 1
}
echo " ✓ Connected"
echo ""
# Step 2: Show remote info
local remote_version auth_mode
remote_version=$(echo "$health_json" | python3 -c "import sys,json; d=json.loads(sys.stdin.read()); print(d.get('version','?'))" 2>/dev/null || echo "?")
auth_mode=$(echo "$health_json" | python3 -c "import sys,json; d=json.loads(sys.stdin.read()); print(d.get('authMode','none'))" 2>/dev/null || echo "none")
echo " Remote OCP v$remote_version (auth: $auth_mode)"
echo ""
# Step 2.5: auto-discover anonymous key from /health (issue #12 §14 Path A).
# The server advertises anonymousKey in /health ONLY when the admin has set
# PROXY_ADVERTISE_ANON_KEY=1 (default off — /health is unauthenticated, so
# advertising exposes the shared key to any LAN-reachable device; issue #109).
# Localhost callers always receive it regardless. When the field is absent,
# ocp-connect falls back to anonymous access / interactive --key (step 3 below).
if [[ -z "$key" ]]; then
local anon_key
anon_key=$(echo "$health_json" | python3 -c "
import sys, json
try:
d = json.loads(sys.stdin.read())
k = d.get('anonymousKey')
except Exception:
k = None
print(k if k else '')
" 2>/dev/null || echo "")
if [[ -n "$anon_key" ]]; then
key="$anon_key"
local _anon_display="$anon_key"
if [[ ${#anon_key} -gt 16 ]]; then
_anon_display="${anon_key:0:8}...${anon_key: -4}"
fi
echo " ⓘ Using server-advertised anonymous key: $_anon_display"
echo " (set by admin via PROXY_ANONYMOUS_KEY; see issue #12 §14 Path A)"
echo ""
fi
fi
# Step 3: Determine if key is needed
if [[ "$auth_mode" != "none" && -z "$key" ]]; then
# Try anonymous access first (zero-config: server may allow it)
if curl -sf --max-time 5 "$base_url/v1/models" >/dev/null 2>&1; then
echo " Server allows anonymous access — no key needed."
echo ""
else
echo " Remote requires authentication."
echo " Ask the OCP admin to run: ocp keys add <name>"
printf " Enter API key (or press Enter to skip): "
{ read -rs key </dev/tty; } 2>/dev/null || key=""
echo
if [[ -z "$key" ]]; then
echo ""
echo " ✗ No key provided — cannot connect to an auth-required remote without a key."
echo " Use: ocp-connect $host --key <key>"
exit 1
fi
fi
fi
# Step 4: Test API access via /v1/models
echo " Testing API access..."
local models_out models_ok=0
if [[ -n "$key" ]]; then
models_out=$(curl -sf --max-time 10 \
-H "Authorization: Bearer $key" \
"$base_url/v1/models" 2>/dev/null) && models_ok=1
else
models_out=$(curl -sf --max-time 10 "$base_url/v1/models" 2>/dev/null) && models_ok=1
fi
if [[ $models_ok -eq 0 ]]; then
echo " ✗ API access failed — key may be invalid or revoked."
exit 1
fi
local model_count
model_count=$(echo "$models_out" | python3 -c "import sys,json; print(len(json.loads(sys.stdin.read()).get('data',[])))" 2>/dev/null || echo "?")
echo " ✓ API accessible ($model_count models available)"
echo ""
# Step 5: Detect shell rc files and OS
local rc_files=()
local is_mac=false
[[ "$(uname)" == "Darwin" ]] && is_mac=true
# Write to all relevant rc files
if [[ "${SHELL:-}" == */fish ]]; then
echo " Note: fish shell detected. Writing to ~/.bashrc — add to fish config manually."
rc_files+=("$HOME/.bashrc")
elif $is_mac; then
# macOS: default shell since Catalina (2019) is zsh.
# Always write ~/.zshrc (create if absent — zsh tolerates an empty file).
# Only write ~/.bashrc if it already exists (don't surprise users with new files).
[[ -f "$HOME/.bashrc" ]] && rc_files+=("$HOME/.bashrc")
# zshrc: always include on macOS; create the file if it doesn't exist yet
[[ -f "$HOME/.zshrc" ]] || touch "$HOME/.zshrc"
rc_files+=("$HOME/.zshrc")
else
# Linux / other: write to whichever rc files already exist or match current shell
[[ -f "$HOME/.bashrc" || "${SHELL:-}" == */bash ]] && rc_files+=("$HOME/.bashrc")
[[ -f "$HOME/.zshrc" || "${SHELL:-}" == */zsh ]] && rc_files+=("$HOME/.zshrc")
# If neither exists, fall back to creating one for the current shell
[[ ${#rc_files[@]} -eq 0 ]] && rc_files+=("$HOME/.${SHELL##*/}rc")
fi
# Step 6: Remove any previously written OCP LAN lines (idempotent) from all rc files
for rc_file in "${rc_files[@]}"; do
if [[ -f "$rc_file" ]]; then
local tmp_rc
tmp_rc=$(mktemp)
if python3 - "$rc_file" "$tmp_rc" <<'PYEOF'
import sys
src, dst = sys.argv[1], sys.argv[2]
lines = open(src).readlines()
out = []
skip = False
for line in lines:
s = line.rstrip('\n')
if s == '# OCP LAN (added by ocp connect)':
skip = True
continue
if skip and (s == '' or s.startswith('export OPENAI_BASE_URL=') or s.startswith('export OPENAI_API_KEY=')):
continue
skip = False
out.append(line)
open(dst, 'w').writelines(out)
PYEOF
then
cp "$tmp_rc" "$rc_file"
fi
rm -f "$tmp_rc"
fi
done
# Step 7: Append new config to all rc files
for rc_file in "${rc_files[@]}"; do
{
echo ""
echo "# OCP LAN (added by ocp connect)"
echo "export OPENAI_BASE_URL='$base_url/v1'"
if [[ -n "$key" ]]; then
echo "export OPENAI_API_KEY='$key'"
fi
} >> "$rc_file"
chmod 600 "$rc_file" 2>/dev/null || true
done
echo " Shell config:"
for rc_file in "${rc_files[@]}"; do
echo " ✓ $(basename "$rc_file")"
done
echo " OPENAI_BASE_URL=$base_url/v1"
if [[ -n "$key" ]]; then
echo " OPENAI_API_KEY=${key:0:8}..."
fi
# Step 7b: System-level env vars (for IDEs, daemons, GUI apps)
if $is_mac; then
# macOS: launchctl setenv makes vars visible to all GUI apps and launchd services
launchctl setenv OPENAI_BASE_URL "$base_url/v1" 2>/dev/null
if [[ -n "$key" ]]; then
launchctl setenv OPENAI_API_KEY "$key" 2>/dev/null
fi
echo ""
echo " System-level (launchctl):"
echo " ✓ OPENAI_BASE_URL set for GUI apps and daemons"
echo " Note: launchctl vars reset on reboot. Add to Login Items or re-run ocp-connect."
else
# Linux: write to environment.d for systemd user services
local env_dir="$HOME/.config/environment.d"
mkdir -p "$env_dir" 2>/dev/null
{
echo "OPENAI_BASE_URL=$base_url/v1"
if [[ -n "$key" ]]; then
echo "OPENAI_API_KEY=$key"
fi
} > "$env_dir/ocp.conf"
chmod 600 "$env_dir/ocp.conf" 2>/dev/null || true
echo ""
echo " System-level (systemd):"
echo " ✓ $env_dir/ocp.conf"
echo " Applies to systemd user services after re-login."
fi
echo ""
# Step 7c: Interactive IDE configuration
configure_ides "$base_url" "$key" "$models_out"
# Step 8: Quick smoke test
echo " Running smoke test..."
# Pick the first available model from /v1/models
local smoke_model
smoke_model=$(echo "$models_out" | python3 -c "import sys,json; d=json.loads(sys.stdin.read()); print(d['data'][0]['id'])" 2>/dev/null || echo "claude-haiku-4-5-20251001")
local chat_payload="{\"model\":\"$smoke_model\",\"messages\":[{\"role\":\"user\",\"content\":\"Reply with OK only.\"}],\"max_tokens\":10}"
local chat_out chat_ok=0
if [[ -n "$key" ]]; then
chat_out=$(curl -sf --max-time 30 \
-H "Authorization: Bearer $key" \
-H "Content-Type: application/json" \
-d "$chat_payload" \
"$base_url/v1/chat/completions" 2>/dev/null) && chat_ok=1
else
chat_out=$(curl -sf --max-time 30 \
-H "Content-Type: application/json" \
-d "$chat_payload" \
"$base_url/v1/chat/completions" 2>/dev/null) && chat_ok=1
fi
if [[ $chat_ok -eq 1 ]]; then
local reply
reply=$(echo "$chat_out" | python3 -c "import sys,json; d=json.loads(sys.stdin.read()); print(d['choices'][0]['message']['content'].strip())" 2>/dev/null || echo "(response received)")
echo " ✓ Smoke test passed: $reply"
echo " Note: smoke test only verifies OCP is reachable and the key is valid."
echo " It does not verify your IDE/agent end-to-end. To verify OpenClaw works,"
echo " restart it (\`openclaw gateway restart\`) and send a test message to your bot."
else
echo " ⚠ Smoke test failed (proxy is reachable but chat completion did not succeed)."
echo " The env vars have still been written. Check the remote OCP logs."
fi
echo ""
echo " Done. Reload your shell to apply:"
for rc_file in "${rc_files[@]}"; do
echo " source $rc_file"
done
}
main "$@"